Вы находитесь на странице: 1из 92

Rural Economic Environment Macro Economic Data The GDP of the Indian Economy was estimated to have grown

at 8.1 per cent during 2003-04 as against 4.0 % in 2002-03 GDP from agriculture sector estimated grew at the rate of 9.1 % instead of least year's rate of (- ) 5.2 per cent during 2002-03 Agricultural production during 2003-04 has been estimated to increase by 19.3 per cent as compared to a decline of 15.6 per cent during 2002- 03 Foodgrains stock in March 2004 was provisionally estimated at 22.8 million tonnes, as against 64.8 million tonnes during June 2002. Select Economic Indicators Sr. No Economic Indicators 2001-02 2002-03 2003-04 (P) 1 Growth in overall real GDP (%) 5.8 4.0 Q 8.2 A 2 Growth in GDP from agriculture & allied sectors (%) 6.5 (-)5.2 9.1 3 Growth in agricultural production (%) 7.6 (-) 15.6 19.3 P 4 Growth in foodgrains Production (%) 8.2 (-) 18.2 21.0 P 5 Growth in industrial production (%) 2.7 5.7 6.9 6 Gross Domestic Savings (as % of GDP) 23.5 P 24.2 Q NA 7 Inflation as measured by WPI (%) 1.6 6.5 4.6 8 Fiscal deficit( as % of GDP) 6.2 5.3 4.6 P 9 Trade Balance (US$ billion) -7.6 -7.5 -13.7 10 Foreign Exchange Reserves (US$ bln.) 51.0 71.9 107.4 11 External Debt (US$ billion) 96 104.6 112.1@

P-Provisional, NA - Not Available, @ as at end December 2003 Sources: (1) Economic Survey, 2003-04, (2) Monthly Review of Indian Economy ,CMIE, February & May, 2004 [Cited in Annual Reports (2003-04) of NABARD] WTO Agreement on Agriculture: Implication on Indian Agriculture The Agreement on Agriculture (AoA), which forms a part of final act of Uruguay Round of Multilateral Negotiations (1986-93) was signed by member countries including India in April 1994 at Marrakech in Morocco. The AoA contains provisions in three broad areas: (i) Market access: There can be no restrictions on farm trade except through tariffs. All the non-tariff barriers such as Quantitative Restrictions (QR) on imports in the form of quotas, import restrictions, import licensing, which were in existence before the Agreement were to be replaced by tariffs on import to provide the same level of protection. In the process of tarrification, however, the tariffs were to be reduced by 36 per cent over a period of 6 years by developed countries and by 24 per cent in 10 years by developing countries. All the member countries had to provide market access opportunities to other member countries to the extent of 3 per cent of their domestic consumption going up to 5 per cent at the end of implementation. (ii) Domestic support: The member countries to club support into two categories, (a) support having no or negligible distorting effect on trade, i.e. Green and Blue Box measures and (b) trade distorting support, i.e. Aggregate Measure of Support (AMS). The Agreement stipulates a reduction of total AMS by 20 per cent by developed countries in 6 years (1995-2000) and 13 1/3 per cent by developing countries in a period of 10 years (1995-2004) taking 1986-88 as the base period. (iii) Export subsidies: The Agreement stipulates that the developed countries reduce the volume of subsidized exports by 21 per cent and budgetary outlay for export support by 36 per cent over a period of 6 years, the base period being 1986-90. For developing countries, this reduction is 14 per cent in volume terms and 24 per cent in terms of budgetary outlay on export subsidies over a period of 10 years. Implications of AoA on Indian Agriculture India has removed QRs for 1429 items by March 2001: (715 by March 2000 and 714 items by March 2001). Of these, 208 items belong to agriculture sector. India has earlier successfully revised the binding levels on 15 tariff lines, which included skimmed milk powder, spelt wheat, etc. India has also not made any commitment to provide minimum market access opportunities. India can also take suitable measures under WTO's Agreement on Safeguards if there is a serious injury to domestic producers due to surge in imports. In India, product specific support is negligible, while non-product specific

support, i.e. subsidies on agricultural inputs such as power, irrigation, fertilizers, etc. is well below the permissible level of 10 per cent of the value of agricultural output. India is, therefore, not under obligation to reduce domestic support extended to agriculture. The export subsidies as listed in the agreement, which attract reduction commitments are not extended in India. However, developing countries like India are free to provide certain subsidies such as subsidizing of export market costs, internal and international transport and freight charges, etc. Agriculture and Rural Sector A. Rainfall during 2003-04 During the South - West monsoons in 2003 (June - September period), the cumulative rainfall of 92.2cm (67.3 cm in 2002) was 2 per cent above the Long Period Average (LPA) of 90.3 cm , with excess or normal rainfall in 33 out of 36 meteorological subdivisions (as compared to only 15 meteorological sub - divisions in 2002) in the country. The rainfall was well balanced , having equitable spatial and temporal distribution, without any prolonged breaks.

B. Area under Major Crops The satisfactory South West monsoons helped increase the area under major crops, viz.,rice, wheat, coarse cereals, pulses, oilseeds, cotton and sugarcane from 142.4 million ha. in 2002-03 to 160.5 million ha. during 2003-04. Area Under Major Crops (million ha.) Major Crops/Year Area 2002 -2003 2003-2004 * % change Rice 35.1 39.4 12.3 Wheat 24.8 27.3 10.1 Coarse Cereals 26.8 29.8 11.2 Total Pulses 22.3

27.2 22.0 Oilseeds 22.0 25.3 15.0 Cotton 7.0 7.4 5.7 Sugarcane 4.4 4.1 (-) 6.8 Total 142.4 160.5 (-) 12.7 * Provisional Source: Monthly Review of Indian Economy, CMIE, October 2003 & April 2004. [Cited in Annual Reports (2003-04) of NABARD] C. Capital Formation in Agriculture During 2002-03, the Gross Capital Formation in agriculture at 1993-94 prices was Rs. 18657 crore. The shares of public sector and private sector in total capital formation were 24.3 per cent and 75.7 per cent respectively. Investments in agriculture as a percentage of GDP declined from 1.6 during 1995-96 to 1.3 during 2002-03.

Gross Capital Formation in Indian Agriculture (at 1993-94 prices) (Rs. crore) Year Total Public Private Per cent share Public Private 2000-01 16906 3927 12979 23.2 76.8 2001-02 17328 4127 13201 23.8 76.2

2002-03 18657 4538 14119 24.3 75.7

*Quick Estimates, Source: Economic Survey 2003-2004 [Cited in Annual Reports (200304) of NABARD] D. Inputs used in Agriculture i. Seeds The Seed Bank Scheme was introduced in 2000 in order to make available quality seeds for meeting contingent requirements. 117 lakh quintals of certified/quality seeds were to be distributed during 2003-04. The production of foundation seeds was estimated at 6.5 lakh quintals and breeder seeds at 0.5 lakh quintals. ii. Fertilizers The domestic production of fertilizers marginally declined during 2003-04 compared to 2002-03. The requirement of fertilizer imports was reported at 2 million tonnes during 2003-04. Fertlizer consumption has been erratic since 2000-01, hovering between 16.1 million tonnes to 17.5 million tonnes in the last four years. Production and Consumption of Chemical Fertilisers (Million tonnes) Year Production Consumption Imports 2000-01 14.7 16.7 2.1 2001-02 14.7 17.4 2.4 2002-03 14.5 16.1 1.8 2003-04 14.2 17.5 2.0

*estimated Source: Economic Survey 2003-04 [Cited in Annual Reports (2003-04) of NABARD] iii. Irrigation Irrigation potential had gone up to about 94.7 million hectares by 1999-2000 (compared to 22.6 million hectares in 1950). This included 59.4 million ha. (63%) through minor irrigation (MI) projects. The Ultimate Irrigation Potential (UIP) has been reassessed at 139.9 million hectares. So far about 68% of potential has been harnessed. The investment in MI structures is expected to create additional irrigation potential of 12 million ha. during the Tenth Plan period. iv. Rural Credit The agricultural credit by Cooperative banks, Commercial Banks, RRBs and other agencies increased from Rs.69,560 crore during 2002-03 to Rs. 86,980 crore during 2003-04. The total credit flow is projected to reach a level of Rs.1,04,500 crore in 200405. E. Agricultural Production The production of foodgrains was estimated at 210.8 million tonnes during 2003-04 as against 174.2 million tonnes during 2002-03. The production of foodgrains and commercial crops over last three years is furnished in the following table: Agricultural production (in million tonnes) Year/Crops 2001-2002 2002-03 2003-04* Rice 93.3 72.7 86.4 Wheat 72.8 65.1 72.7 Coarse Cereals 33.4 25.3 36.8 Pulses 13.4

11.1 14.9 Total Foodgrains 212.9 174.2 210.8 Oilseeds 20.7 15.1 25 Sugarcane 297.2 281.6 244.8 Cotton** 10 8.7 13.5

*Third Advance Estimates, ** Million bales of 170 kg. Each. Source:Economic Survey, 2003-04 [Cited in Annual Reports (2003-04) of NABARD] F. Agricultural Trade i. Agricultural Exports The share of export of agriculture and allied products in the total exports declined from 18.5 per cent during 1990-91 to 13.1 per cent during 2002-03.The share of agricultural exports to total exports of the country came down to 11.7 per cent during April February 2003-04 from 12.9 per cent during the same period in 2002-03. The export performance as at the end - January 2004, revealed that export of cereals, groundnut, other oilseeds, fresh fruits/ vegetables, floriculture, dairy and poultry products increased, while the export of some of the agro and agro - processed commodities, viz., tea, spices, tobacco, basmati rice, processed fruits and juices, etc., declined. ii. Agricultural Imports The imports of agricultural products increased to $2.8 billion in 2003-03 from $ 2.3 billion in 2001-02. The share of agri- imports to total merchandise imports increased from 4.5 %to 4.6%, over the same period (Table 8.24). Edible oil which accounted for almost two - thirds of the total agri- imports, is the single largest item of agri- imports into the country. The dominant share of this item continued to persist even in 2003-04.

Imports of pulses are also significant and accounted for 20-30% of agri- imports, although its share declined to 14% in 2003-04. G. Procurement and Stock of Foodgrains An estimated, 36.5 million tonnes of grains were procured during 2003-04. The stock of foodgrains as on March 2004 was estimated at 22.8 million tonnes as against a peak level of 64.8 million tonnes. I. Animal Husbandry and Dairy Sector India owns the largest livestock population in the world, accounting for nearly 57 per cent of the world buffalo population and 16 per cent of the cattle population .India was the largest producer of milk in 2002-03 with a total production of 86.4 million tonnes. During 2002-03 the livestock sector accounted for about 5.4 per cent of the total GDP and 27.7 per cent of the GDP from the agricultural sector. Progress in Production and Availability of Milk Years Milk Production (Million Tonnes) Per Capita Availability (Grams/Day) 1950-51 17.0 124 1960-61 20.0 124 1970-71 22.0 112 1980-81 31.6 128 1990-91 53.9 176 2000-01 80.8 215 2001-02 84.3 221 2002-03* 86.4 226

*provisional Source: Economic Survey 2003-04 Livestock sector plays an important role in the national economy and in the socio economic development of the country. Livestock sector produces 86.4 million tonnes of milk, 41.7 million tonnes eggs, 52.1 million kgs of wool and 4.94 million tonnes of meat. H. Fisheries India is the sixth largest producer of fish in the world and second in inland fish production. The fishery sector has also emerged as an important source of employment generation. The fishery sector contributes 1.1 per cent of the total GDP and 5.4 per cent of the GDP from agriculture sector. Fish production reached a level of 6.2 million tonnes in 2002-03. I. Agricultural Reform Initiatives Various states have come up with reform initiatives to strengthen their agriculture sector in general and agro- processing, marketing and infrastructure development in particular. a) Agricultural Marketing In order to facilitate private participation in agricultural marketing , the GoI circulated a draft model scheme of the " Agricultural Produce Marketing , Development and Regulation Act, 2003" to enable state governments to make suitable amendments in their respective Agriculture Produce Marketing Acts. Karnataka, Madhya Pradesh and Punjab have taken the lead in allowing private participation in agricultural marketing. b) Irrigation and Water Supply In order to ensure water conservation and management, reforms have been introduced in the irrigation sector by many states. Andhra Pradesh has enacted the Andhra Pradesh Land, Water and Trees Act, 2002 under which 50 per cent subsidy would be provided subject to a ceiling of Rs. 50, 000 for installation of micro irrigation systems. ' Pani Panchayats' were encouraged by the Government of Orissa. Karnataka amended the Karnataka Irrigation Act in 2000, enabling the formation of 1712 Water Users' Associations and authorising them to raise resources for maintenance of irrigation structures. Punjab and Jharkhand have also taken initiatives to provide irrigation facilities to farmers. c) Agro- processing and value addition

Jharkhand in its industrial policy provided special thrust to agro- processing industries. Under an MoU signed with APEDA , Agri- Export Zones (AEZs) were established in three districts to boost vegetable production.The Madhya Pradesh Krishi Upaj Mandi Abhiniyam, 2003 was amended to enable contract farming agreements. Govt. of Uttar Pradesh announced a policy package to boost the agro - based food processing industry.In Maharashtra, AEZs are being set up for mango, onion, orange, pomegranate and banana. Punjab is promoting agro- exports by providing subsidised packaging and refrigeration at both production and consumer ends. Andhra Pradesh aims to provide incentives in terms of different production and processing related subsidy schemes, concessional electricity Tariff, exemption from payment of market cess, etc. In Karnataka, five Food and Agri Technology Parks (FATPs) with complementary services are being established for the benefit of the agri-business community. d) Forest and Tribal Development The Govt. of Andhra Pradesh has issued comprehensive orders for improvement of degraded forests through the participation of local people by the introduction of ' Joint Forest management Programme'.In Kerala, a 'New Basic Needs Package' for tribals with three components, viz., health insurance, quality education and housing and sanitation has been announced for the poor tribals.

HIGHLIGHTS 2004-2005 Credit Operations Short-term credit limits sanctioned during 2004 - 05 For SCBs, RRBs - seasonal agricultural operations - Rs.10185.06 crore For RRBs - other than seasonal agricultural operations - Rs.216.83 crore. For SCBs - financing Weavers' Cooperative Societies- Rs.349.89 crore. Long term loans sanctioned to 7 State Governments for contribution to the share capital of co-operative credit institutions aggregated Rs.32.98 crore. Liquidity support to SCBs - Rs.1914.24 crore Liquidity support to RRBs - 158.78 crore Investment Credit to CBs, SCARDBs, SCBs, RRBs and other eligible institutions - Rs. 7605.29 crore. Kisan Credit Card Scheme During the year ( upto Feb 2005), 70.43 lakh cards issued by co-operative banks, RRBs and commercial banks. Since inception in 1998-99, 4.84 crore cards issued. Rural Infrastructure Development Fund

GoI announced Rs. 8000 Crore for RIDF XI ( 2005-06 ) As at the end of March 2005, RIDF sanctions under all the tranches of RIDF amounted to Rs. 42948.51 crore against which the disbursements were Rs. 25384.02 crore. SHG bank linkage programme - Highlights- 2004-05 During the period April 2004 to March 2005 - 5,39,385 new SHGs were financed by banks to a tune of Rs 29.94 billion by way of loans. Cumulatively, banks have lent Rs 68.98 billion to 1,618,476 SHGs. 35,294 branches of 560 banks (Commercial banks- 48; Regional Rural banks-196; & Cooperative banks 316) situated in 563 districts in the 30 states of the country are participating in the programme.( Data- provisional) About 24.25 million poor households have gained access to formal banking system through SHG bank linkage programme. Nearly 90% of the groups are women only groups. Promotional grant assistance Grant assistance extended by NABARD to various agencies/ institutions for promotion & linkage of self-help groups during the year as well as cumulatively is given below; Agency During 2004-05 Cumulative as on 31.3.2005 Number Amount (Rs million) For promotion & linkage of SHGs Number Amount (Rs million) For promotion & linkage of SHGs NGOs 263 42.66 24234 1048 193.87 139513 RRBs 12 2.97

3890 93 30.55 38935 DCCBs 26 10.63 12560 55 23.03 28110 RRB- Regional Rural Bank; DCCB- District Central Cooperative Bank; Capacity building initiatives Around 42,812 bank officials, 4,246 NGO staff, 7,063 government officials and 2,07,916 self help group members trained with grant support from NABARD. In addition, about 161 faculty members of various banks' training establishments were also trained. Cumulatively 1,016,600 persons trained through various SHG related capacity building programmes. Financial Highlights NABARD was set up under an Act of Parliament, NABARD Act, 1981 and has a capital base of Rs.2000 crore contributed by RBI and Government of India. As on 31 March 2004, the highlights are as under: Total Assets of Rs.55889 crore of which owned funds are Rs.22660 crore. Net profit before tax stood at Rs.1460 crore. A high Capital Adequacy Ratio of 39.41% as against a minimum of 9% stipulated by RBI. NPA as low as 0.0014% of advances as on 31 March 2004. Summary of Balance Sheet (Rs. Crore) As on March 31 1999-2000 2000-01

2001-02 2002-03 2003-04 Capital 2000 2000 2000 2000 2000 Reserves & Surplus 15184 16922 18266 19469 20660 Borrowings & Deposits 15212 18966 23495 27634 31588 Others 971 928 1338 968 1641 Total 33367 38816 45099 50071 55889 Refinance and Loans & Advances 30175 35772 41063 45361 48790

Investments 1833 1388 1298 1355 2412 Others 1359 1656 2738 3355 4687 Total 33367 38816 45099 50071

Rural Development of India My Quote: Happyness is the key of our life. So always keep smiling face (to make others happy). Rural development is my passion and I would like to start my own project very soon. Currently I am studying the various peoples' experiences in this regard. I am badly looking for like minded for joint endeavours. Simultaneously I am also looking for a girl who is interested in similar activities, to marry. How can we develop??: After discussing with many many people (across the globe), I came to understand that : Main issues for rural development are: 1) Education (everyone should be educated "irrespective of age") 2) Roads construction (also a major concern) 3) Self-Employment generation 4) Health awareness 5) Giving awareness of opportunities availabilities for them 6) Family Planning

Other important issues are: Wastemanagement, Alcohol (and tobocco) prohibition, organic farming, drainage management etc. What I want you to do??: Any of the following way to make collaboration if you have similar interests. 1) Please mail me with your interests to malapati@ google's mail (gmail.com) 2) Join orkut community (http://www.orkut.com/Community.aspx?cmm=1344880) NOTE: Joining in ORKUT requires an invitation (I can send an invitation if you send me an e-mail with your interests). What about financial resources??: This is what currently I am also thinking about. I am not interested in making corpus funds. According to requirements (as and when need arises), collection of money from individuals/organizations may be a good idea. Ofcourse this is where I assume that we are dedicated young people who is not considering rural development as a career-oriented but as an obsession. When will it be started??: After finishing the study of various other people experiences. This may take few months to years. At this moment I am with several models. I am now observing how well they were successful in others' cases(rather which areas/conditions they may suit). Do you wish to appreciate our effort?: We strongly request you to be associated with us. If you do not have much time, you may keep in touch with us. Sometimes moral support itself may encourage us to progress further. Is it an established organization??: Not yet. We have plans to start soon. Stay in touch with us. How I got interested??: As early as, I was just 8 years old, I joined in 6th grade/class to become the youngest pupil in my class (and could not mix up with much elder people even for playing). And also due to my father's "too much" restrictions, I did not like to interact with classmates/any student but used to spend 3 hrs a day in the village Library (reading (moral) story books & novels). This helped me to get much more interest in the social life of different regions of people. Also my father used to ask me to byheart all moral poems

of telugu (use of this I realized now). Once I came across a photograph accedentally (about 7 years before) which made to rethink about my career. 1 11 MARCH 2004 CAN WE DISCERN THE EFFECT OF GLOBALIZATION ON INCOME DISTRIBUTION? EVIDENCE FROM HOUSEHOLD SURVEYS Branko Milanovic 1 World Bank, Development Research Group and Carnegie Endowment for International Peace ABSTRACT The effects of globalization on income distribution within rich and poor countries are a matter of controversy. While international trade theory in its most abstract formulation implies that increased trade and foreign investment should make income distribution more equal in poor countries and less equal in rich countries, finding these effects has proved elusive. The paper presents another attempt to discern the effects of globalization by using the new data derived directly from household surveys. It looks at the impact of openness (proxied by the trade/GDP ratio) and direct foreign investment on relative income shares across the entire income distribution. We find strong evidence that at low average income level, it is the rich who benefit from openness. As income level rises, the situation changes and it is the relative incomes of the poor and the middle class that rise compared to the rich. Openness seems to make income distribution worse before making it betteror put differently the effect of openness on income distribution depends on countrys average income level. Keywords: income distribution, inequality, globalization JEL classification: D31, F15, I3. 1 Email: bmilanovic@ceip.org. I am grateful to Prem Sangraula, Dimitri Kaltsas and Gouthami Padam for excellent research assistance. The paper has been much improved thanks to the comments of James K. Galbraith, Aart Kraay, Peter Rundell, Martin Ravallion, Alan Winters and three anonymous referees. I am also grateful to the participants of the Conference on Globalization and Inequality held at the Brookings Institution in Washington in June 2002, and Massachusetts Avenue Development Seminar held at the Center for Global Development in Washington in November 2003 for comments and suggestions. The paper was written as a continuation of the Research Project 684-84 financed by the World Bank Research Grant. The views expressed in the paper are authors own, and should not be attributed to the World Bank or its affiliated organizations. 2 1. Introduction: Globalization and its effect on income distribution The issues of globalization and income inequality have during the last ten years received

a huge attention. Most of it was concentrated on the effects of globalization on withincountry inequality in rich economies. In other words, the discussion was mostly on how globalization is (or might) affect wage and income inequality in the United States or Western Europe (e.g. Slaughter and Swagel 1997; Dluhosch, 1998; Schott, 1999; Lejour and Tang, 1999). A second strand of the literature was focused on how globalization might affect world or international income distribution principally via differences in mean per capita growth rates between the countries (see Milanovic, 2004; Milanovic and Yitzhaki, 2002; Melchior et al. 2000; Schultz 1998, Sala-i-Martin 2002). Only recently has there been more interest in how globalization affects within-country distribution in less developed countries (Cornia and Kiiski, 2002; Lustig and Kanbur, 1999; Ravallion 2001; Galbraith and Kum 2002). There is also a discussion of the effects of globalization on poor countries growth and technology transfer (Gundlach and Nunnenkamp, 1999). There are theoretical models of income distribution as affected by trade (e.g. Wood, 1994, and 2000; Benarroch and Gaisford, 1997; Kremer and Maskin, 2003). The detailed empirical analyses of the effects of economic change, including market reforms and increased international integration, on within-country income distribution are limited mostly to Latin American countries. Harrison and Hanson (1999) and Robertson (2000) study wage inequality in the wake of Mexican trade reforms; Beyer, Rohas and Vergara (1999) look at the similar issue in the context of Chile; Arbache (1999) studies the effect of market liberalization on intersectoral wage dispersion in Brazil; Behrman, Birdsall and Szekely (2003) assess the impact of various policy changes (including trade liberalization and capital account opening) on wage differentials in Latin America. But there are relatively few papers that try to assess the impact of openness on income distribution in both poor and rich countries. This is the objective of this paper. Relevant to this approach are two recent papers by World Bank researchers. In both cases, the main objective was to look at the relationship between openness and growth, but both papers provide some 3

interestingeven if conflictingevidence on the relationship between openness and inequality. Lundberg and Squire (1999; see also the more recent version 2003) consider growth and inequality to be simultaneously determined. They find, in an unbalanced panel comprising more than 700 observations from 125 countries and covering the period 1960-1998, that openness, measured by the Sachs-Warner (0-1) indicator, has either no effect or a mild negative effect on income growth of the bottom quintile. However the greater the mean income of the country the more significant and positive is the effect of openness on income growth of the poor quintile (Lundberg and Squire, 1999, pp. 31-32).2 The implication of the Lundberg-Squire results is that the effect of openness on income distribution varies in function of level of development (countrys mean income). This is very similar to the results obtained by Barro (2000) and Ravallion (2001, p. 1811). They also find statistically significant non-linearity in the relationship between openness and inequality, with openness associated with increased inequality in poor countries. In a somewhat different twist, Spilimbergo, Londono and Szekely (1999, p.88) control for countries endowments in skilled labor, capital and land. They find that in capital-rich countries, openness reduces inequality while in countries with the abundant skilled labor openness increases inequality. They argue that the former effect is driven by the reduction of capital rents once domestic capital markets open up while the latter effect is consistent with the Heckscher-Ohlin framework. A different conclusion is reached by Dollar and Kraay (2002; see also an earlier version Dollar and Kraay, 2000). They also use an unbalanced panel covering the same period and (almost exactly) the same countries as Lundberg and Squire, and find first, that openness (defined as exports plus imports as a share of GDP3) is positively associated with per capita income growth, and second, that this effect carries across all income quintiles.4 Trade has no systematic impact on inequality. The implication of their finding is that trade is neutral to 2 They also find that openness is a trade-off variable: its effect is positive for growth, negative for equality (the result obtained when using the Gini coefficient to measure inequality). 3 With GDP measured in PPP termsa feature which, as we shall find out, has some important implications. 4 There are some definitional differences compared to Lundberg and Squire. For example, Dollar and Kraay space

observations on the mean income of the poorest quintile by at least five years. They do so in order to avoid relying in their estimation on too many annual and adjacent income distribution statistics from the rich countries. In addition, it is reasonable to be concerned with the medium-term effects of growth on inequalitya fact which would be obfuscated by overreliance on annual data. 4 income distribution, and since trade is good for growth, the effects across all income groups are positive and the samewhere the same means that each deciles gain is proportional to its initial income. (In other words, the rich benefit more in absolute amount, but not in relative terms.) In a similar vein, Birdsall and Londono (1997 and 1998), report no differences between growth in income of the poorest and other quintiles due to trade variables, though initial distribution of land and education do matter. Finally, Li, Squire and Zou (1998) have, in one of the sensitivity runs of their main model, export-toGDP ratio (a proxy for openness) as an explanatory variable of the Gini coefficient. They also find no statistically significant effect of openness on the Gini coefficient. These different findingsespecially because they are derived by using very similar data sourceshave generated intense discussion. Dollar and Kraay (2000, pp. 16-18) address some empirical and methodological differences between their paper and that by Lundberg and Squire. A recent paper by Ravallion (2001) attempts to find out where the difference in results comes from, and to reconcile their findings. Birdsall and Hamoudi (2002) argue that Dollar and Kraay fail to distinguish between the effects of negative growth spells on the poor vs. others, and that volatility including spells of negative growth is more likely in more open economies.5 Thus, in conclusion, we have inconsistent results regarding the effects of openness on inequality. On the one hand, Li, Squire and Zou (1998), Birdsall and Londono (1998) and Dollar and Kraay (2001, 2002) find that openness has no systematic and significant effect on inequality. On the other hand, Lundberg and Squire (1999), Barro (2000) and Ravallion (2001) find that openness has a negative effect on equality in poor countries, and moreover, in some of the formulations, that it has a negative effect on real income of the poor as well. The

conclusions thus run the full gamut, from openness having a negative effect on real income of the poor, to raising income of the poor less than income of the rich in relative terms, to raising both the same (in relative terms, again). Note however, that there are no results that show openness reducing inequality, that is raising real income of the poor by more (in percentage terms) than income of the rich. Let alone raising absolute incomes of the poor by more. 5 The latter point is also made by Easterly, Islam and Stiglitz (2001). 5 The new data base The objective of this paper is to provide some additional empirical evidence on how globalization affects income distribution in developed and developing countries using the newly developed data base created in the context of the work on world income distribution. The advantages of World Income Distribution (WYD) data base are twofold. First, it is entirely based on national household surveys anchored around three benchmark years (1988, 1993 and 1998) so that income inequality data are almost fully mutually comparable; that is, the data are all derived from nationally representative surveys and use the same definition of recipient units (individuals) who are ranked by per capita household income (or expenditure).6 Second, the data do not provide one or two synthetic inequality measures (e.g. Gini coefficient or Theil index) but give the actual data on income levels across ten deciles of income distribution. It thus describes practically the entire income distribution which none of the papers reviewed above has been able to do. This ability to look at what is happening behind a change in one summary statistic is crucial if we want to get a better grasp of the effects of globalization since it is thought to affect different parts of distribution differently. The total number of available surveys with decile data is 321 with 95 countries (surveys) in 1988, and 113 countries both in 1993 and 1998. Almost 2/3 of deciles are calculated from individual (micro) level data.7 There are 82 countries (called common sample countries) with decile data for all three years. Incomes are expressed in $PPP (international dollars) of each benchmark year but since we are interested in withincountry distributions, the currency used is irrelevant (the share of a decile is the same whatever

currency we use). However, the use of PPP data is relevant if we want to look at the impact of openness in function of level of development, i.e. mean country income. 6 There are problems though. Some surveys are income- and some expenditure-based. However, since countries generally specialize in producing either one of the other type of survey, it is mostly the cross-over countries (those that in one year use income and in another year expenditure survey) that pose problems. Their number however is limited. There are 9 such cases out of 113 surveys in 1993, and 13 such cases (out of 113 surveys) in 1998. Overall, about 1/3 of all available distributions are expenditure-based. In order to simplify the writing, we shall always refer to income distributions though. There are other problems too. We cannot be sure that income and expenditures are defined equally in all the surveys (e.g. imputation practices for home consumption or services of consumer durables differ, fringe benefits may or may not be included ) and underreporting, particularly of property incomes, is rife in both poor and rich nations. Household surveys thus tend to show income distributions less unequal than they are. 7 The proportions differ between the years though. In 1988, the percentage is 44; in 1993, it is 55, and in 1998, it is 70. 6 The coverage of world income and population by the WYD data is shown in Table 1. Additional details regarding the data sources and surveys are given in Milanovic (2004, forthcoming, Chapters 9 and 10) and Milanovic (2002, Appendix 1). The data are available at www.worldbank.org/research/inequality/data. Table 1. Coverage of world GDP and population by household survey data (in percent) Population GDP (in US$) 1988 1993 1998 1988 1993 1998 Africa 48.0 76.1 67.1 48.7 85.2 71.2 Asia 92.5 94.9 94.4 94.4 93.2 95.6 E. Europe/FSU 99.3 95.2 100 99.4 96.3 100 LAC 87.4 91.8 93.0 90.2 92.8 95.2 WENAO 92.4 94.8 96.6 99.3 96.2 96.3 World 87.3 92.4 91.6 96.5 95.4 96.0 Source: Milanovic (2004; forthcoming). WENAO is Western Europe (inclusive of Israel), plus Northern America, and Oceania (more exactly, Australia and New Zealand only). LAC is Latin America and the Caribbean. The paper is organized as follows. In Section 2, we look at the definition of globalization

and what are the channels through which it may affect within-country income distributions. In Section 3, we propose a simple model on how globalization affects income distribution. Section 4 gives some descriptive statistics of the data with an emphasis on the measures of globalization. Section 5 gives the estimates of the regressions. Section 6 summarizes our findings and gives some policy implications. 2. What it means to be globalized? It is sometimes useful to begin with the official definition. The official World Bank definition of globalization is Freedom and ability of individuals and firms to initiate voluntary economic transactions with residents of other countries. Empirically globalization translates into greater mobility of the factors of production (capital and labor) and greater world integration through increased trade and exchange of ideas. Several recent papers that compare the two globalization waves, the one at the end of the last century up to 1914 and the current one, look precisely at these indicators (Bordo, Eichengreen, and Irwin, 1999; Williamson, 1996; Craft, 2000; Baldwin and Martin 1999): how much trade there is now (as the share of world GDP) compared to a century ago, how much direct and portfolio foreign investment, and how easy it is 7 for people to move or to settle in different countries. The studies come with a mixed verdict on the past vs. current globalization. Portfolio investments and trade as a share of world GDP are about the same now as then, direct foreign investment is greater as is the ease of travel, but the ability to resettle elsewhere is less. Thus, it appears that both labor and capital are in some sense more, and in some sense less, mobile than they were a century ago, and that trade is about as important now as it was then. However, our objective here is to look at how thus defined and empirically understood globalization affects income distribution. . Consider first the effects of openness on less developed countries. These countries are affected principally in two ways. First, they are able to export more of their own goods (and to import more), and second, they can be expected to be the recipients of direct foreign and

portfolio investments from the capital-rich countries. According to the simple version of the Heckscher-Olin-Samuelson (HOS) model, less developed countries will tend to export low-skillintensive products (because low-skill labor is their abundant factor and its price is low). Foreign investors will also tend to invest in low-skill intensive processes in poor countries. Moreover as the more advanced countries have an advantage in skill-intensive products and specialize in production and export of these, there should be also a reduction in relative wages of highlyskilled workers in less developed countries. When we translate this into what it should imply for income distribution, and approximate the latter by the ratio between high-skill and lowskill wage, it appears that income inequality within the less developed countries should go down. Mirroring these developments, income distribution in more developed countries should become more unequal. This is directly derived from factor price-equalization theorem in its most abstract formulation (see Freeman, 1995 and caveats therein), and is argued, for example in Wood (1995, 1998) and Tang and Wood (1999). Moreover, as less developed countries continue their process of modernization which implies improvement in educational attainment, the relative supply of high-skill workers increases compared to low-skill workers (although not to the extent that it would reverse the comparative advantage of the country). This seems to further reduce the wage difference between the high- and low-skilled workers and to shrink wage (and thus income) distribution. In conclusion, relative demand shifts occasioned by globalization would tend to favor lessskilled workers in poor countries, and so do relative supply shifts brought about by better educational 8 achievement. The reverse would, of course, hold for rich countries where globalization would tend to favor high-skilled workers although that effect should, in the longer-term, be lessened by their greater supply. Income inequality in poor countries should unambiguously decrease in both the short- and the long-run while it might increase in the short-run in rich countries (before the pro-inequality effects are not counterbalanced by greater supply of better educated workers). What may be the offsetting elements? There are, at least, two. First, rather than looking at

globalization through HOS lenses, we may look at it as a Kuznetsian process. Suppose that instead of two types of labor (low- and high-skill) we have three types of labor (low-, medium-, and highly-skilled). Globalization may produce movement of labor from low-wage sector (agriculture) where wage differentiation is minimal, to medium-skill sectors (in urban areas) where wage differences are larger. Then, even if the ratio between the top and bottom shrinks (that is, the ratio between the average high-skill and the average low-skill wage becomes smaller), overall wage (and hence, we assume for the moment) income inequality might increase simply because of the greater wage differentiation in the middle. In conclusion, the ratio between the average wages of different types of labor is not sufficient to describe what happens to the distribution. In effect, as soon as we move from the very simplifying HOS assumptions of two types of labor, the effects of trade (and globalization) on income distribution in developing countries become ambiguous. In a model developed by Wood (1994), there are three types of labor (skilled, intermediate and unskilled). Then, very poor countries that open up may experience increased inequality because demand generated by openness helps those with basic and high education (that is those with intermediate and high skills) and reduces the income share of those with no education. This introduces not only greater realism in the assumptions, but highlights the fact that speaking of a single effect of globalization on developing countries may be wrong. Poor developing nations with the abundance of unskilled agricultural labor may experience an increase in inequality (e.g. Bangladesh), while those that enter globalization with a mostly educated labor force, and where primary education is the norm (e.g. South Korea) might see increased trade bring about less inequality. 9 The second offsetting element may be the following. Although wages constitute, even in developing nations, the largest chunk of total income, there are two other income sources that affect income inequality significantly, and both of them are likely to move in direction of greater inequality. The two income sources are self-employment income (including homeconsumption),

and capital (property) income. The share of self-employment income would tend to go down as people move from subsistence agriculture (this assumes that peasants are mostly landowners) to becoming wage-workers. Since self-employment income and home consumption in particular tend to be relatively equally distributed, their decreasing share will add to inequality. The importance of capital income will depend on the distribution of wealth and on what happens to the real interest rate, whose level is, in turn, dictated by what happens in rich countries. This is particularly so in an era of globalization and more or less free movement of capital. Since property income is strongly concentrated among top income classes, that element too might provide a strong countervailing force to decreasing inequality. 3. Channels of influence on the entire income distribution and estimation issues By definition, absolute income level of the i-th decile in the j-th country at the time t can be written as a function of an inequality index (Ijt) and mean income of the country (mjt).8 ) , ( jt jt ijt m I f y = (1) The relative income of the i-th decile (normalized by the mean) is then9 ) ( jt jt ijt I g m y = (2) We then assume that level of the inequality index depends on the levels of the variables listed below. 8 Deciles go from the poorest, 1, to the richest, 10. 9 The movement from (1) to (2) implies the homogeneity assumption 10 (1) two standard globalization variables, namely openness (OPENj) measured as the sum of exports and imports in countrys GDP, and direct foreign investment as a share of GDP (DFIj), (2) financial depth (FDj), the ratio of M2-to-GDP, introduced on the assumption that greater financial depth should reduce the importance of the financial constraint to borrow for education purposes, and thus should help those who are talented but lack resources (see, for example, Li, Squire and Zhou, 1998), and (3) an indicator of democracy (DEMj), introduced on the assumption that democratization, through the median voter hypothesis, should lead to a greater redistribution and reduction in inequality (see Milanovic, 2000 and the literature review in Gradstein and Milanovic, 2004). We need to explain the use of openness as a measure of globalization because it was

criticizedalthough in a somewhat different contextby a number of authors (Rodrik 2000; Birdsall and Hamoudi 2002; Lubker, Smith and Weeks 2002). There are two key critiques. First, that openness represents an outcome that governments cannot influence, and not a policy, or choice, variable like, for example, level of tariff protection. Often times openness is presented the critiques argueimplicitly at least, as a policy. Yet the trade ratio may decline not because the country follows a more closed policy but because of balance of payments difficulties as was the case when the terms of trade for commodity producers collapsed in the early 1980s (see Birdsall and Hamoudi, 2002). If one is not aware of this, one may ascribe growth slowdown in the 1980s to decreased openness whereas the true cause lies in an exogenous shock. Second, the trade/GDP ratio is often treated as a determinant of growth whereas the true causality may be reversefrom growth to trade. Both of these criticisms are, I think, valid, but they do not affect the use of openness in our context, as a variable to explain inequality. The reason is as follows. Here, we are not concerned with policies, that is, with whether a country follows open policies or not, nor with the growth-trade causality. Our concern is solely how a given quantum of trade regardless of how it is achieved, whether by open or closed policiesaffects distribution of income. In other words, we do not take openness to be a choice variable but just take openness as is, and ask whether it has some discernable impact on income distribution Moving to the other variables, financial depth and democracy are not thought to be linked directly with globalization even if one might plausibly entertain such a view too. For example, one can regard increasing financial depth, that is increasing monetization of the economy, to 11 proceed directly from better integration of a country into the international economy, and democratization to occur in response to greater international exchanges. However, we view these two variables as controls for the non-globalization related part of the influence on income distribution, and orthogonal to the globalization-proper variables. We introduce them primarily to avoid misspecification of the model. We rewrite (2) in the reduced form as

) , , , ( jt jt jt jt jt ijt DEM FD DFI OPEN m y = (2a) In this formulation, there is no role for income as an explanatory variable. The reason is as follows. The argument that income affects inequality and should hence be included on the RHS (right-hand side) is based on some variant of the Kuznets-type relationship. Whether one subscribes to the Kuznets hypothesis or not, it is clear that income serves only a proxy for several structural changestransfer of labor from a more equally distributed agriculture to a more unequal industry, or educational change (increasing share of the highly skilled people and decreasing education premium), or demographic change (increasing share of the elderly and rising social transfers). These structural changes are all associated with a rising GDP per capita. Once we solve in for income by introducing its structural correlates like financial deepening and democracy, there is no additional independent role played by income. However, we need also to take into account the possibility that the globalization variables will not affect the share of a given decile the same regardless of countrys level of development. Consider the following fact. Increased openness and direct foreign investments will, as the theory tells us, tend to benefit low-skilled workers in poor countries since it would be the lowskillintensive industries which would be both attractive to foreign investors and likely to take advantage of export opportunities. Thus, we would expect the signs of OPEN and DFI variables to be positive among the bottom deciles in poor countries (they increase bottom deciles income shares). For a rich country, the situation is exactly the reverse. Openness will mean that it is the low-skilled workers in rich countries that would be exposed to increased foreign competition, and we would expect that the signs of OPEN and DFI variables would be negative in a rich country setting. The coefficients of the two globalization variables will therefore vary in 12 function of the income level of the country. Ideally, of course, the coefficients should vary in function of the skill composition of each income decile and countrys income level. However,

since we do not have information on who exactly is in each decile and what is the skill composition of people per decile, we shall use countrys income level to interact with the openness variable. Interaction between OPEN and income has been used in at least several papers: by Barro (2000), Ravallion (2001), and Dollar and Kraay (2002). We can then write (omitting time subscripts) for each decile: eij DEMj i FD i m DFIj i DFI i m OPENj i m OPEN i i mj yij jjjjjij+++(+++++=21076543)*)*( (2b) The coefficients vary across deciles and are thus subscripted. We expect the signs of i6 and i7 associated with respectively financial depth and democracy to be positive among the low deciles, and negative among the higher income deciles. There are ten pooled cross-section regressions: one for each income decile run across all countries with the same independent variables. The regressions such as (2b) can be run independently (with one omitted) or as a simultaneous system (seemingly unrelated regressions) with a constraint. 10 The constraint ensures that the sum of coefficients adds up to zero: this is necessary since an increase in the share of some deciles must be balanced by the decrease in the share of other deciles. Because of shares likely autocorrelation (within country and across years), the regressions are run with robust (Huber/White) standard errors. There are two additional problems: (i) endogeneity, and (ii) robustness of the results to the introduction of other variables. 10 If we assumed that the slopes are homogeneous across countries and that intercepts are fixed (different between countries), we could use fixed-effect (FE) estimator. The advantage of the FE (or firstdifference) estimator lies in the fact that it allows us to argue that marginal effects of openness (and other explanatory variables) are the same across countries while letting inequality be determined (through varying intercept) by other unobservable countryspecific effects. This seems quite reasonable. The problem however is that our panel is very short (three observations only) and that since shares, within county, change but slowly most of the data variability is contained in cross-sectional observations. Thus, the use of the FE estimator yields very poor results. 13 The endogeneity problem may plague both openness and other RHS variables. Inequality

might influence financial depth, or democracy, or government spending (variable introduced below). To adjust, in part, for it, all RHS variables are calculated as five-year lagged averages. There is also a substantive reason for it too: to reflect the fact that openness or financial depth do not affect income distribution instantaneously. Time is required to do so. But we also address endogeneity more fully by instrumenting the possibly endogenous variables by their lagged values and other instruments, and using GMM (generalized method of moments) estimator whose efficiency properties are superior to those of the traditional IV/2SLS estimators.. The robustness of the results can be questioned because our RHS variables may not include all relevant variables that affect income shares. We augment our regression (2b) by adding the extent of government spending (as a share of GDP) and real rate of interest. The first is expected to be pro-poor, the second, due to the typically high concentration of assets in the hands of the income-rich, to be pro-rich. We introduce both of these controls to check for the robustness of the results. In addition, one of the strong results from the inequality literature is that there are regional patterns in income distribution, namely that Latin American or African countries tend ceteris paribus to have unequal income distributions (see Fields, Higgins and Williamson 1999, Fields 2001, p.67) and Asian countries relatively equal (see Milanovic, 1994). We shall therefore control for the regional effects as well.11 Broadening the range of the control variables addresses also another potential source of endogeneity, namely having inequality and RHS variables be determined jointly by another omitted variable. As we increase the number of controls on the RHS, we basically tend to cover most (or ideally all) the bases for such an effect. 11 I am grateful to a referee for this point. 14 4. Descriptive statistics Before trying to link globalization and other macro variables to changes in income distribution, we need to define the variables more precisely. For the distribution, we use the data

on annual per capita $PPP incomes of each decile from 321 surveys and 129 countries in total (with 82 countries being a balanced panel). As mentioned before, the income data are for the benchmark years 1988, 1993, and 1998. Each decile contains 10 percent of individuals (not households) of a country. The dependent variable is defined as the ratio between deciles own income and overall mean income. All RHS variables are calculated as the averages over the five-year periods. There are two reasons for this rather than simply using a single value for 1988, 1993 and 1998. First, the distribution data are only benchmarked in 1988, 1993 and 1998. The actual surveys which we use to calculate the decile data might have been conducted in the years around 1988 (say, 1986 or 1989).12 Thus the averaging which exists for the dependent variable is accompanied by a similar averaging of the controls. . Second, even if all the surveys were conducted in the same year, there would be some advantage in relating changes in income shares to several years average share of exports and imports in GDP. This in order to avoid having the results being swamped by noise, that is very short run changes which cannot have much influence on a sluggish variable like income distribution. As mentioned before, globalization is reflected in two variables: openness share of combined exports and imports in GDPand the share of direct foreign investments in GDP of the recipient country. Thus openness that is associated with income distribution around 1988 is taken to be the average of exports and imports over GDP during the five-year period ending in 1988 (that is, 1984-88). Likewise, openness that is associated with income distribution in 1993 and 1998 is defined as the average over respectively 1989-93 and 1994-98 periods. Identical calculations are done for other RHS variables. 12 Overall, however, more than 70 percent of surveys are within a year of the benchmark; more than 90 percent of surveys are within two years of the benchmark date. 15 Table 2 shows mean-normalized average incomes of each decile in 1988, 1993 and 1998. For example, we see that in 1988, on average (calculated across all countries)13 the bottom

deciles income was 30.7 percent of the mean. The same decile calculated across the commonsample countries received an income equal to 30.3 percent of the mean. By 1993, the bottom deciles income was only 23.5 percent of the mean in one case and 24.4 percent in the other. Finally, in 1998, it declined even further to 23.3 percent of the mean. Note that between 1988 and 1993, relative incomes of the bottom eight deciles went downwith the negative change the largest among the poor decileswhile the relative income of the top two deciles went up, again with the greatest positive change among the very top. The situation changed between 1993 and 1998. All deciles between the second and the seventh (inclusive) gained, while the very bottom decile and the three top deciles lost (all of course in relative terms). Table 2. Mean-normalized average incomes of each decile (across countries, not weighted for population) Note: Deciles formed based on per capita income or expenditures obtained from household surveys. The decile ratio is the ratio between the average income of the tenth and the first decile. Table 3 shows the recent upsurge in globalization as reflected in the openness variable. It shows the increase in the combined share of exports and imports in GDP, all at current prices. There is a sustained increased in the (unweighted) share of openness from around 70 percent in the mid-1980s to more than 90 percent at the turn of the century. The (dollar) weighted share of trade in world GDP similarly increased from 38 to 44 percent (see Table 4). The higher 13 Each country is one observation regardless of its population size. All countries Panel (common sample countries) 1988 1993 1998 1988 1993 1998 First 0.307 0.235 0.233 0.303 0.244 0.233 Second 0.441 0.375 0.380 0.437 0.391 0.387 Third 0.539 0.476 0.482 0.535 0.495 0.491 Fourth 0.635 0.571 0.581 0.631 0.593 0.590 Fifth 0.736 0.677 0.686 0.733 0.701 0.697 Sixth 0.855 0.804 0.810 0.853 0.831 0.821 Seventh 1.000 0.959 0.962 1.000 0.984 0.972 Eighth 1.201 1.182 1.181 1.202 1.207 1.188 Ninth 1.541 1.566 1.552 1.548 1.580 1.553 Tenth 2.745 3.156 3.138 2.757 2.973 3.068 Total 1 1 1 1 1 1 Number of countries 95 113 113 82 82 82

Decile ratio 8.9 13.4 13.5 9.1 12.2 13.2 16 unweighted trade/GDP average ratio reflects the fact that smaller (and poorer) countries trade shares are greater. It is notable that consistently the most closed economies are Brazil and Japan. The most open economy is almost throughout Singapore. Table 3. Share of combined exports and imports in GDP (unweighted; cross country) Year Number of countries Average share of openness (in percent; all countries) Minimum (in %) Maximum (in %) Average share of openness (in percent; common sample countries) 1985 124 72.8 13 (Lao) 317 (Spore) 70.7 1986 125 67.5 10 (Iran) 308 (Spore) 66.0 1987 128 68.1 10(Sudan) 341(Spore) 67.0 1988 129 69.1 15(Sudan) 375(Spore) 69.5 1989 130 73.5 13(Sudan) 362(Spore) 71.6 1990 132 76.5 14(Brazil) 539(Suriname) 73.3 1991 130 75.0 15(Brazil) 399(Suriname) 72.6 1992 140 75.9 16(Tajik) 385(Suriname) 70.1 1993 150 76.8 16(Japan) 326(Spore) 72.6 1994 152 79.7 16(Haiti) 331(Spore) 75.7 1995 154 82.8 16(Brazil) 339(Spore) 80.2 1996 155 83.9 15 (Brazil) 328(Spore) 83.1 1997 155 84.9 17 (Brazil) 317(Spore) 85.0 1998 153 86.8 17 (Brazil) 457(Eq. Guinea) 86.0 1999 152 85.3 19(Japan) 313(Spore) 85.1 2000 149 91.9 20(Japan) 341(Spore) 94.0 Source: Own calculation from World Development Indicators; World Bank. SIMA (Statistical Information Management and Analysis) database, World Bank. The number of common sample countries is 82. The increase in openness was registered in all the regions (Table 4). The increase was very

large in transition economies, Latin America and Asia. It is remarkable that differences in openness between the continents are relatively small, and that the view, often expressed, of insufficient integration of Africa in global economy is belied by these numbers.14 Openness of Africa is not much different, or is even higher, from that of the rest of the world. In 2000, for example, Africas trade-to-GDP ratio was 62 percent, some 15 percent higher than that of the WENAO (rich) world (not shown in the table).15 Africas low share in total world trade simply 14 For example, Yusuf (2003, p. 68) in an article on globalization points to inability of African countries to integrate with world economy. 15 This is the dollar-weighted openness (the ratio between total value of Africas trade and its total GDP). 17 mirrors its low share in total world income. Or differently put, the problem with Africa may not be trade but small size of domestic markets. For 85 countries whose openness increased between 1985 and 2000, the average unweighted change was 27 GDP points. The most significant increases were registered by Malaysia (127 percentage points), Angola (107), and Hong Kong (87). For 32 countries where openness decreased, it did so by an average of 15 GDP points. The most significant decreases were for Mauritania (52), Bahrain (49), and the Bahamas (48). Openness for several of the largest (by GDP) countries went up. For example, for the US it increased from about 16 to 24 percent of GDP, for China, from 23 to 49, for India, from 14 to 30, from Brazil, from 19 to 23 percent, for Germany from 42 to 67 percent. But, on the other hand, for Japan, openness went down from 24 to 20 percent of GDP. Table 4. Openness (exports plus imports) as percentage of GDP (unweighted regional averages unless stated otherwise) First period (1984-88) Second period (1989-93) Third period (1994-98) Change in openness

Africa 62.9 68.4 73.2 +10.3 Asia 72.1 80.8 91.7 +19.6 Latin America 57.8 72.2 78.7 +20.9 E. Europe/FSU 64.0 61.6 84.6 +20.6 WENAO 71.0 69.3 75.7 +4.7 World 70.4 75.8 83.5 +13.1 World (dollar weighted) 37.7 38.8 43.9 +6.2 Number of countries 130 150 155 Source: Own calculation from World Development Indicators; World Bank. SIMA Database, World Bank. The openness ratios for each period are calculated as (A) the means of five-year averages of all the countries (this is exactly the same definition as used when creating the variables for the regression analysis below). These values are slightly different than if (B) the period openness ratios were calculated as the means of all individual countries openness ratios for these years. The reason is that with the method (A) each countrys five-year average counts the same (each country provides one five-year average), while with the method (B) countries whose trade ratios are available for all years will count more than countries whose trade ratios are available in (say) only three out of five years. Even more dramatic were increases in foreign direct investments as percentage of GDP of the recipient countries. The unweighted importance of foreign direct investments increased from less than 1 percent of GDP in the late 1980s to 4.6 percent in 2000 (Table 5). The increase was 18 most dramatic in the second part of the 1990s (see Table 6) when the importance of DFI measured in terms of total world output almost doubled. If we compare the amounts of DFIs in 1985-90 and ten years later (1995-2000), for 62 countries the share of DFI inflows in GDP increased by an average of more than 3 GDP points, while for only seven countries DFI became less important. In ten countries (Lesotho, Ireland, Bolivia, Sweden, Panama, Denmark, the Netherlands, Chile, Czech republic, and Bulgaria) the share of direct foreign investment in GDP in the most recent period exceeded by more than 5 GDP percentage points their share in the late 1980s. For China, the importance of DFI went up,

over the same period, from an average of 0.5 percent of GDP to between 3 and 4 percent of GDP. India, which started with almost no direct foreign investments, reached some of one percent of GDP in the late nineties. In the US, similarly, the share went up from 0.5 percent of GDP to between 2 and 3 percent. As we have recently come to expect (Lucas 1990), per capita DFI (in current dollar terms) and GDP per capita (also in current dollar terms) are positively associated with each 10 percent increase in income accompanied by a little over 10 percent increase in received foreign investments. 16 16 The hypothesis of unitary elasticity is accepted: income elasticity was somewhat greater in the first than in the last five-year period. 19 Table 5. Foreign direct investment as percentage of recipient countrys GDP (unweighted average) Year Number of countries Percentage of GDP Maximum a/ 1985 65 .70 5.7 (New Zealand) 1986 65 .67 4.2 (New Zealand) 1987 66 .84 3.5 (New Zealand) 1988 67 .86 4.2 (Lesotho) 1989 68 1.17 7.9 (Nigeria) 1990 70 1.25 6.2 (Zambia) 1991 70 1.17 8.1 (Malaysia) 1992 80 1.13 8.8 (Malaysia) 1993 82 1.39 7.5 (Malaysia) 1994 82 1.79 8.3 (Nigeria) 1995 82 2.28 29.5 (Lesotho) 1996 83 2.39 30.5 (Lesotho) 1997 83 3.22 26.1 (Lesotho) 1998 83 4.13 29.7 (Lesotho) 1999 81 4.17 24.6 (Sweden) 2000 81 4.63 24.3 (Ireland) a/ Luxembourg, which in all years has the highest share of direct foreign investment in GDP is not shown. Source: Own calculations from UNCTAD Handbook of International Trade and Development Statistics, 1996, 1997, 2000. 20 Alike trade, the flow of direct foreign investment has increased in all regions of the world

with the most significant unweighted increases occurring in Africa, Latin America, and transition economies (Table 6). 17 Table 6. Direct foreign investment as percentage of GDP (unweighted regional averages unless stated otherwise) First period (1984-88) Second period (1989-93) Third period (1994-98) Change in DFI Africa 0.9 1.3 3.8 +2.9 Asia 0.5 1.2 1.6 +1.1 Latin America 0.7 1.4 3.5 +2.8 E. Europe/FSU 0.0 0.6 3.0 +3.0 WENAO 1.1 1.4 2.2 +1.1 World World (dollar weighted) 0.8 0.7 1.2 0.8 2.8 1.4 +2.0 +0.7 Number of countries 67 82 88 Source: Own calculations from UNCTAD Handbook of International Trade and Development Statistics, 1996, 1997, 2000. The period averages are calculated as explained in Table 4. We are less interested in the other control variables, financial depth (M2/GDP), democracy, and income. The former is measured in a straightforward fashion, as the ratio of M2 to GDP (see Table 7). The dramatic decline in financial depth in transition economies is due both to the very high level of money to GDP ratio before the transition, 18 and then to the effects of hyperinflation which reduced real money balances. Democracy is measured by the Democracy variable from PolityIV database created by Monthy Marshall, Keith Jeggers, and Ted Gurr.19 The variable ranges from 0 (absence of democracy) to 10 (best). Finally, income is, as mentioned

before, mean income as calculated from household surveys and expressed in 1995 international dollars. 17 In dollar-weighted terms, the picture is somewhat different. For example, in the fiveyear period ending in 1998, the unweighted DFI/GDP was 3.8 percent for Africa and 2.2 percent for WENAO (see Table 6); but the GDPweighted percentages were respectively 1.7 and 1.4. 18 This was known as the money overhang problem: too much cash chasing too few (price-controlled) goods. There was some recovery in monetization however in 1999 and 2000 (not shown in Table 7). 19 The data are available at www.cidcm.umd.edu/inscr/polity/. Democracy is defined as general openness of political institutions. 21 Table 7. M2 as percentage of GDP (unweighted regional averages) First period (1984-88) Second period (1989-93) Third period (1994-98) Change Africa 47 38 37 -10 Asia 68 82 78 +10 Latin America 30 31 33 +3 E. Europe/FSU 76 33 28 -48 WENAO 60 63 65 +5 World 53 50 48 -5 Number of countries 66 76 80 Source: Own calculations from World Development Indicators, World Bank (SIMA database). The period averages are calculated as explained in Table 4. Democracy variable shows a sustained increase in democracy in all the regions over the three periods. The most important gains were registered in Eastern Europe and the former Soviet Union. Population-weighted democracy, largely held back by the absence of progress (as measured by the variable) in China, also improved although less than unweighted democracy. Table 8. Democracy proxied by the Polity IV Democracy variable (unweighted regional averages unless stated otherwise) First period (1984-88) Second period

(1989-93) Third period (1994-98) Change Africa 1.1 1.5 3.0 +1.9 Asia 3.0 4.2 5.0 +2.0 Latin America 5.5 7.0 7.8 +2.3 E. Europe/FSU 0.1 4.2 6.0 +5.9 WENAO 9.5 9.7 9.8 +0.3 World 3.5 4.8 5.9 +2.4 World (weighted by population) 3.8 4.5 5.0 +1.2 Number of countries 129 129 130 Source: Own calculations from Polity IV database. Democracy index ranges from 0 (least democratic) to 10 (most democratic). For explanation on how the index is derived, see Polity IV available at www.cidcm.umd.edu/inscr/polity/. The period averages are calculated as explained in Table 4. 22 5. Estimation of the regressions We estimate ten level regressions for each formulation, with the first formulation being the most parsimonious one ij j j j j j i j j j e DEMj i FD i m DFIj i DFI i m OPENj i m OPEN i i m yi + + + ( + + + + + = 2 1 0 7 6 5 4 3 ) * ) * ( (2b) where all the variables and coefficients are already explained, and subscript t is omitted for simplicity. All the RHS variables are five-year averages (except mean income which is contemporaneous). The results of the simultaneous decile estimation (seemingly unrelated regressions or SURE) are shown in Tables 9-10. The results of instrumental variable GMM estimation are shown in Tables 11-13. The first regression is an unbalanced panel run across 201 decile shares in years 1988, 1993 and 1998.20 As can be seen in Table 9, for all the deciles between the second and fifth, increased openness negatively affects their income shares. However, the negative effect of openness is lessened for richer countries as the interaction term between openness and mean income is positive. Openness would therefore seem to have a particularly negative impact on the

poor and middle-income groups in poor countrieswhich is directly opposite to what we would have expected based on standard theory. It is only when income level (calculated from household surveys) reaches over $4,000 in purchasing power terms, that is around the income level of Malaysia and Brazil, that for the poor and the middle classes openness becomes a good thingthat is, raises their share in total income. 20 The total number of surveys is, as mentioned before, 321. We lose some of them because of lack of other RHS variables. 23 How large is the openness effect? Let us suppose that we are looking at a poor country whose mean income is $PPP 2,000 per capita, and whose second deciles share of total income is about 4 percentage points (an average value in our sample). The second deciles mean income is therefore $PPP 800. Let now the trade/GDP ratio increase from 0.7 to 0.9 (again, an average change between 1985 and 2000). This will reduce the decile's share of the total pie to about 3.9 percent, or absolute income to $PPP 780 (of course, absent any other effect including change in total income).21 For the top two deciles, openness exerts the opposite effect. It reduces the income share of the rich and does it the more the higher the overall mean income. On balance, therefore, openness is pro-equality in rich countries and pro-inequality in poor countries. Direct foreign investments have a negative effect on the income share of the poor in rich countries (the income and DFI interaction is negative; that is, the higher the income, the more the negative effect on the share), but otherwise they are not statistically significant. Financial depth, as we would expect, increases income share of the poor and middle class, while democracy has no significant effect. The regressions in Table 10 are the same as in Table 9 except that we now add real rate of interest and government expenditures as a share of GDP among the explanatory variables.22 The main thrust of the results is unaffected except that democracy now becomes positively related to the shares of the four middle deciles and direct foreign investment play no role whatsoever. R2 21 This is obtained as follows. At $PPP2000, the sensitivity of the decile ratio variable is -0.09 + (0.00002 * 2000) =

-0.05 which means that, with an infinitesimal increase in openness, the decile ratio will go down by 0.05, that is from 0.4 to 0.395. In other words, the share of the second decile will decrease from 4 percent to 3.95 percent of total income. If we now let openness to increase from 0.7 to 0.9, the effect will be (0.9-0.7) * 0.05 = 0.01 or 0.1 percentage point . So, the effect of a 20 percentage point increase in the trade share will be decline of 0.1 percentage point in the deciles income share. 22 The nominal interest rate is the deposit rate on 12-month deposits as reported in IMFs International Financial Statistics (various issues; the variable is 60L...ZF). The real rate is obtained by deflating the nominal by the 12month consumer price index (also as reported in International Financial Statistics). The data are available on World Bank SIMA. The data on government expenditures are the sum of central government (consolidated accounts), local and state or provincial governments expenditures. The data are taken from the IMFs Government Financial Statistics available on-line in World Bank SIMA database. 24 increases significantly to an average value of 0.5 as both real rate of interest and government expenditures enter as strong predictors of decile shares. Interest rate is shown to be statistically strongly anti-poor: it reduces the share of the deciles between the second and the eighth and raises that of the top two. How strong is this effect? Income share of the top decile is about 30 percent (of total country income). Each percentage point increase in real rate of interest raises that share by a little over 0.1 percent. In other words, real income of the rich (assuming the size of the pie to be fixed) goes up by 1/3 of one percent. Since (1) the period 1988-98 was characterized by rising income inequality (see Table 1) and (2) there is a pro-rich effect of high interest rates, it is interesting to look at changes in real interest over the same years. The unweighted average real interest rate across countries in our sample shows a mild increase. For example, during 1996-2000, the average real interest was always positive ranging between less than 1 percent and more than 3 percent p.a.; but during 1985-90 period, the rate was negative in three out of six years. The number of countries with

negative real deposit interest rates is also smaller now although it is not negligible (there were 17 out of 69 countries with negative real interest rates in 2000). The role of rising real interest rates as a factor in increasing inequality since the early 1980s seems not to have received as much attention as it deserves. The introduction of real interest and government expenditures in the regressions pushes the turning point of the effect of openness on decile shares past the PPP$7,000 mark where countries such as Spain and Israel are located.23 In other words, openness becomes pro-poor only at higher average income levels. Government expenditures have a positive effect on the shares of all low and middleincome deciles, and a negative effect on the shares of the two top deciles. Thus, government expenditures play a role directly opposite to that of high real interest rates. A ten percentage point increase in government expenditures/GDP ratio raises the bottom deciles share of the pie (total countrys income) by 0.35 percent. This represents more than one-tenth of what the bottom decile receives on average (see Table 2). 23 This is mean per capita income as calculated from household surveys. 25 Democracy affects positively income shares of the middle deciles. This is an interesting result which suggests that earlier works which have failed to detect the effect of democracy on inequality (Bollen and Jackman, 1985; Gradstein, Milanovic and Ying, 2001, but for an exception see Tavares and Wacziarg, 2001) might have done so because democracy affects primarily income shares of the middle groups while leaving unchanged those at the top and the bottom. In consequence, synthetic inequality measures like the Gini coefficient may not show much change. In Table 11, we show the results of GMM estimates. Openness and government expenditure as a share of GDP which may be thought endogenous are instrumented by their lagged values and countrys population. The Hansen J statistics (test of overidentifying restrictions) is throughout insignificant indicating that instruments are valid. R2s remain high and most of the conclusions are unchanged. Openness alone reduces income share of all deciles between the second and the sixth. The interaction term between openness and income is again positive for these deciles, and the turning point occurs at somewhat higher income level, at

around $PPP8,000. Government expenditures and financial depth are pro-poor as before, and high real rate of interest is anti-poor. Democracy is significant for the middle deciles. Direct foreign investments do not matter individually or interacted with income. In Table 14, we add two more controls: regional dummies (with WENAO as the reference category) and the five-year average rate of inflation. The general quality of the results improves, and R2 goes up to a rather high average value of 0.7. The role of openness becomes sharper: now it reduces the share of the bottom seven deciles, raises that of the top two. The interaction term between openness and income is statistically significant almost throughout. The turning point occurs around $PPP 8,000. Government expenditures are still pro-poor, and real rate of interest is pro-rich and a significantly strong predictor of income shares across the entire distribution. As expected from the literature, inflation is strongly significant, and affects income shares of the poor and the middle class negatively. The regional dummies for Latin America and Eastern Europe/FSU are significant almost throughout, the first being anti-poor, the second propoor (compared to the omitted WENAO dummy variable). Both Africa and Asia show lower shares of the middle deciles, and higher share of the very top one. Once we introduce regional dummies, mean incomea proxy for the level of developmentbecomes significant and affects income 26 shares of the poor negatively, and those of the rich positively. Thus, poorer countries would tend to have (everything else being the same) more unequal distribution. At the end, we look at the sensitivity of our results to different definitions of openness, and compare them to those of Dollar and Kraay (2001 and 2002). We have used several alternative definitions of openness: trade/GDP in constant US dollars converted at market exchange rates (i.e. volume of trade variable kopen given in Penn World Tables 6.1), and exports/GDP ratio in current prices. Both formulations yield the same results as the use of our preferred measure

(trade/GDP in current dollars) and for the reasons of space we do not report them here. The results however can be obtained from the author. Perhaps more interesting is to explore the difference between our results and those reported by Dollar and Kraay (2002). Their much-quoted paper has two very important conclusions: first, that mean growth and the growth rate of the bottom quintile display, on average, a unitary elasticity (meaning that poors percentage income increase is the same as the average increase), and second, that openness does not significantly affect the income share of the bottom quintile. The second finding is of obvious relevance for this paper, since our results do differ from those of Dollar and Kraay. The difference may be due to a difference in the sample (theirs is larger), period that is being covered (Dollar and Kraays data go back to the 1970s), or the inequality measure used (ours is better because the deciles we use are almost all calculated directly from individual household surveys while Dollar and Kraays quintile data come from a very heterogeneous collection of sources with some of them being extrapolations). Yet probably the most important difference resides in the definition of the openness variable. In this paper, openness is defined as trade/GDP with both expressed in nominal dollar terms. In Dollar and Kraay, openness is defined quite interestingly and rather unusually as the ratio between trade (in 1985 dollars) and GDP also in 1985 prices but measured in international dollars (see Dollar and Kraay 2002, p. 220). It is a consistent definition since both the numerator and the denominator are in international dollars (trade is by definition conducted at international prices), but it is a different indicator from (say) volume of trade as given in Penn World Tables. There trade is also in the numerator (in the same 1985 prices) but GDP is calculated in exchange rate dollars of the same year. The difference is important because Dollar and Kraay by using PPP dollars to express 27 the denominator (GDP) will significantly increase it for poor countries and thus reduce the importance of trade in their GDP. In Table 12 we show the results of the previous regression except that the Dollar-Kraay

measure of openness is substituted for ours.24 The results do change. Openness is no longer significant for the bottom decile while it affects negatively deciles two to five. Unlike our measure, it does not show a positive effect for the share of the top two deciles. The interaction between openness and income remains positive for the bottom deciles, and negative for the top. It would thus seem that openness reduces the share of the richest decile at any mean income level (that is, in poor, middle income and rich countries). But if we look at openness and its effect on the bottom two deciles, it is not unambiguously negative at low average income levels as in Table 11. Here however for the very bottom decile openness exerts a positive impact. Why is there a difference between the results obtained by these two measures of openness? To see this, consider how the they behave when trade in poor countries expands.25 The effect of the increase in trade on the Dollar-Kraay measure will be small or nil since the bulk of these countries GDP will still consist of non-tradables that are valued at (high) international prices. But both trade/GDP in current prices and trade/GDP in constant prices (and at market exchange rates) may go up significantly. A very nice example is provided by India and China. In nominal terms, trade/GDP ratio has increased in India from 14 percent in 1985 to 30 percent in 2000, and in China over the same period from 23 percent to 49 percent. In volumes (given by Penn World Tables), the increase has been from 28 to 53 percent of GDP for China, and from 18 to 25 percent for India. In PPP terms used by Dollar and Kraay, however, the trade ratio over the same period has barely budged from 6 to 7 percent in India, and went up from only 7 to 9 percent in China. A stark illustration of the difference implied by the use of different measures is shown in Figure 1: the lowest line one is always the Dollar-Kraay measure, the one in current prices (broken line) and the one in constant prices move almost in unison. 24 We have only slightly changed the calculation of the Dollar-Kraay measure by expressing both trade and GDP in 1995 (rather than in 1985) international prices. 25 There would be no important difference between the two measures for the rich countries becsause their GDPs calculated at PPP or mareket exchange rate are quite similar

28 Figure 1. Different measures of openness for India and China, 1980-2000 India China Note: The Dollar-Kraay measure (trade-over-GDP both in 1995 international dollars) in always the lowest line; trade/GDP in current prices is given by the broken line; trade/GDP in constant prices (solid line) is volume of trade from Penn World Tables 6.1 (variable kopen). Consider now the situation where income inequality rises simultaneously (as indeed it did in both countries). The Dollar-Kraay measure will fail to detect much of a relationship between the openness and inequality because it is artificially sluggish. In our case however an increase in openness will be associated with greater inequality. Which approach makes more sense? In this contextwhere we are interested in how important is international trade for income creation and income distribution in a given countryit is trade/GDP ratio in nominal prices which matters. The role that trade plays in total income of a country, that is, in peoples earnings, depends on how much actual income is generated in trade compared to purely domestic activities. Income distribution is affected by the actually received incomes, not by notional incomes that are ascribed through imputation of international prices to domestic goods and services. If I am a barber in India, for my income, and for the income distribution in India, it is my actual local currency pay that matters, not my total output valued at international prices. Or similarly, going back to the example of China, surely exports which account for 26 percent of nominal GDP in the year 2000, play in total peoples .05 .1 .15 .2 .25 .3 Trade/GDP ratios 1980 1985 1990 1995 2000 year .1 .2 .3 .4 .5 Trade/GDP ratios 1980 1985 1990 1995 2000 year 29 income a role that is commensurate with that share rather than with 5 percent that exports represent in the PPP-calculated Chinas GDP. 26 In conclusion, the results of the level regressions show that when a country is relatively poor, increased openness raises the income share of the top, and reduces the income share of the

poor groups as well as of the middle class. (We are throughout talking of shares, not absolute incomes.) However, at some medium- to high level of mean country income, between $PPP4,000 and $PPP7,000 per capita depending on the specification, the income shares of the poor and the middle class begin to be positively affected by openness while the income share of the rich begins to decline. Finally, for the rich countries, openness is associated with an increasing share of the bottom and middle deciles, and a decreasing share of the top deciles. Openness thus makes inequality chart an inverted U shape as income level increases. At low income levels, openness is bad for equality; at high income level it promotes equality. This suggests that only the middle-income countries may behave as the rigorous version of the trade theory would imply. But poor countries whose equality should be helped by openness, and the rich countries where openness should increase income differentials, behave in the exactly the reverse fashion from what we would expect. These results are only partly consistent with those posited by Wood (1994), or Kremer and Maskin (2003). In both models, poor countries that open up their economies may experience increased inequality because openness helps those with middle and high level of education, but reduces the income share of those with no education. Since they are the largest and the poorest category of people, income inequality goes up. In Kremer and Maskin (2003), workers with no education are basically unemployable by rich countries firms because of their low productivity (they cannot be matched with more skilled workers), and hence globalization marginalizes them. It is only when basic education becomes the normand even the poor have itthat openness exerts an incomeequalizing effect. This is what we might be picking up in the results which show that at some middling level of income, the share of the lower and middle income classes begins to rise. In other words, a strategy based on exports of manufactures that require at least basic education would be 26 Clearly the greater the PPP value of GDP, the better off is an average citizen since she obviously can consume more goods and services. This type of PPP comparison is useful if we want to compare average welfare levels in two countries. But this is not our objective here.

30 equitable in Korea but inequitable in Burkina Faso or Pakistan (Wood 1994, quoted in Kanbur, 1998). However, in Woods and Kremer and Maskins models, rich countries too exhibit increased inequalitya result we do not find here however. 27 27 To be more exact, Kremer and Boskin (2003, p.17) are agnostic about the effect on inequality in rich countries: inequality may go up or down 31 Table 9. Explaining mean-normalized decile incomes (1988, 1993, 1998): Parsimonious formulation (regressions estimated simultaneously; dependent variable: decile mean income/overall mean income) First Second Third Fourth Fifth Sixth Seventh Eighth Ninth Tenth Open5 -0.0679 -0.0908 -0.0879 -0.0886 -0.0797 -0.0621 -0.0379 0.0104 0.0682 0.437 (0.08) (0.033) (0.037) (0.030) (0.037) (0.078) (0.217) (0.704) (0.133) (0.081) Mean income -0.0000006 0.000002 0.000003 0.000004 0.000004 0.000005 0.000006 0.0000009 0.000003 -0.00003 (0.96) (0.64) (0.526) (0.465) (0.38) (0.247) (0.149) (0.064) (0.544) (0.282) Open5*mean 0.00002 0.000023 0.00002 0.000022 0.000021 0.000017 0.000011 0.000001 -0.000019 -0.00011 income (0.01) (0.002) (0.003) (0.002) (0.002) (0.007) (0.041) (0.852) (0.019) (0.009) DFI5 0.0004 -0.0009 -0.0044 -0.0045 -0.0055 -0.0069 -0.007 -0.0054 0.0010 0.0333 (0.94) (0.886) (0.516) (0.495) (0.38) (0.229) (0.163) (0.221) (0.891) (0.417) DFI5*mean -0.000004 -0.000004 -0.000003 -0.000003 -0.000003 -0.000002 -0.000001 0.0000002 0.000002 0.00002 income (0.011) (0.015) (0.063) (0.046) (0.059) (0.140) (0.365) (0.853) (0.161) (0.074) M2gdp5 0.0853 0.092 0.0813 0.0693 0.0577 0.0500 0.0391 0.0184 -0.0042 -0.4913 (0.004) (0.004) (0.010) (0.024) (0.045) (0.059) (0.090) (0.369) (0.901) (0.09) Democr5 -0.0024 0.0004 0.0006 0.0008 0.0008 0.0007 0.00002 -0.0014 -0.0021 0.0029 (0.33) (0.98) (0.814) (0.754) (0.731) (0.742) (0.992) (0.417) (0.488) (0.858) Constant 0.2396 0.3606 0.4617 0.5628 0.6695 0.7902 0.9457 1.1634 1.5544 3.2522 (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) No of obs 201 201 201 201 201 201 201 201 201 201 "R-sq" 0.179 0.258 0.276 0.293 0.304 0.306 0.274 0.126 0.077 0.284 Note: Coefficients that are statistically significant (at 5 percent level) are shaded. The pvalues between brackets. Suffix 5 indicates five-year average. 32 Table 10. Explaining mean-normalized decile incomes (1988,1993, 1998). Adding real rate of interest and government expenditures as share of GDP (regressions estimated simultaneously; dependent variable: decile mean income/overall mean income) First Second Third Fourth Fifth Sixth Seventh Eighth Ninth Tenth Open5 -0.1191 -0.153 -0.1469 -0.1409 -0.1222 -0.0943 -0.0583 0.0180 0.1241 0.6937 (0.004) (0) (0) (0.001) (0.001) (0.008) (0.070) (0.558) (0.017) (0.007)

Mean income -0.000009 -0.000008 -0.000006 -0.000005 -0.000003 -0.0000008 0.000001 0.000006 0.000007 0.00001 (0.083) (0.152) (0.220) (0.290) (0.430) (0.748) (0.866) (0.138) (0.218) (0.550) Open5*mean 0.000013 0.000018 0.000017 0.000016 0.000014 0.000010 0.000006 -0.000004 -0.0000203 -0.00007 Income (0.068) (0.014) (0.022) (0.020) (0.030) (0.092) (0.289) (0.476) (0.027) (0.106) DFI5 -0.0024 -0.0042 -0.0072 -0.0073 -0.0079 -0.0091 -0.0091 -0.0069 -0.0011 0.0552 (0.702) (0.519) (0.264) (0.243) (0.175) (0.097) (0.064) (0.141) (0.889) (0.161) DFI5*mean -0.000001 -0.000001 -0.000001 -0.000001 -0.0000007 -0.0000006 -0.0000001 0.0000004 0.000002 0.000005 Income (0.382) (0.441) (0.726) (0.627) (0.632) (0.799) (0.934) (0.694) (0.503) (0.743) M2gdp5 0.1376 0.1484 0.1357 0.1234 0.1120 0.1047 0.0889 0.061 -0.0231 -0.8914 (0.001) (0) (0.001) (0.001) (0.002) (0.002) (0.004) (0.036) (0.638) (0) Democr5 -0.0012 0.0029 0.0045 0.0057 0.0063 0.0065 0.0058 0.0035 -0.0004 -0.0339 (0.669) (0.335) (0.123) (0.045) (0.019) (0.010) (0.010) (0.101) (0.91) (0.061) Rint5 -0.0011 -0.0017 -0.0019 -0.0019 -0.0019 -0.0019 -0.0017 -0.0012 0.0024 0.0111 (0.09) (0) (0) (0) (0) (0) (0) (0) (0) (0) Expgdp5 0.2753 (0) 0.3493 (0) 0.3429 (0) 0.3285 (0) 0.2981 (0) 0.2426 (0) 0.1677 (0) 0.0436 (0.288) -0.1862 (0.007) -1.8597 (0) Constant 0.169 0.2591 0.3524 0.4468 0.5528 0.6815 0.8540 1.0984 1.5854 4.0017 (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) No of obs 137 137 137 137 137 137 137 137 137 137 "R-sq" 0.346 0.507 0.534 0.551 0.562 0.552 0.503 0.275 0.224 0.524 Note: Coefficients that are statistically significant (at percent level) are shaded. The pvalues between brackets. Real rate of interest expressed in percent per annum (e.g. 2 percent), government expenditures expressed as a share of GDP (e.g, 0.3 not 30 percent). Suffix 5 indicates five-year average.

33 Table 11. Explaining mean-normalized decile incomes (1988,1993, 1998) (GMM/IV estimation; dependent variable: decile mean income/overall mean income) First Second Third Fourth Fifth Sixth Seventh Eighth Ninth Tenth Openness5 -0.1266 -0.1641 -0.1628 -0.1574 -0.1345 -0.1037 -0.0642 0.0078 0.0979 0.7504 (0.003) (0) (0) (0.0) (0.001) (0.008) (0.056) (0.784) (0.048) (0.005) Expgdp5 0.2823 0.3638 0.3582 0.3431 0.3127 0.2538 0.1768 0.0469 -0.1963 -1.9492 (0) (0) (0) (0) (0) (0) (0) (0.137) (0.002) (0) Mean income -0.000009 -0.000008 -0.000008 -0.000007 -0.000005 -0.000003 0.0000004 0.000004 0.000005 0.000028 (0.093) (0.113) (0.095) (0.091) (0.133) (0.337) (0.862) (0.130) (0.389) (0.233) Openness5*mean income 0.000013 0.000018 0.000018 0.000018 0.000015 0.000011 0.000006 -0.0000028 -0.00001 -0.00007 (0.048) (0.017) (0.021) (0.014) (0.019) (0.062) (0.227) (0.583) (0.056) (0.063) M2gdp5 0.1539 0.1802 0.1582 0.1353 0.1137 0.1007 0.0836 0.0558 0.0147 -0.9149 (0) (0) (0) (0) (0.01) (0.001) (0.002) (0.032) (0.737) (0) DFI5 -0.0020 -0.0051 -0.0085 -0.00911 -0.0101 -0.0109 -0.0102 0.0069 0.0010 0.0665 (0.671) (0.340) (0.154) (0.152) (0.113) (0.092) (0.107) (0.212) (0.845) (0.098) DFI5*mean income -0.0000001 -0.0000009 0.0000002 -0.0000004 -0.0000003 -0.00000001 -0.0000003 -0.0000004 0.000001 0.0000008 (0.318) (0.509) (0.886) (0.803) (0.857) (0.994) (0.848) (0.771) (0.430) (0.934) Democracy5 0.00073 0.0045 0.0067 0.0081 0.0087 0.0084 0.007 0.0040 -0.0017 -0.0442 (0.833) (0.163) (0.044) (0.013) (0.005) (0.003) (0.007) (0.130) (0.684) (0.041) Rint5 -0.0013 -0.0017 -0.0017 -0.0017 -0.0017 -0.0016 -0.0014 -0.0009 0.0024 0.0098 (0.008) (0) (0.001) (0.001) (0.001) (0.001) (0.003) (0.011) (0.023) (0) Constant 0.1578 (0) 0.2348 (0) 0.3339 (0) 0.4329 (0) 0.5414 (0) 0.6756 (0) 0.8519 (0) 1.1073 (0) 1.6193 (0)

4.062 (0) Hansen J 1.959 (0.161) 2.163 (0.141) 2.494 (0.114) 2.384 (0.122) 1.98 (0.159) 1.064 (0.302 0.179 (0.67) 0.103 (0.74) 0.867 (0.35) 0.786 (0.375) No of obs 138 138 138 138 138 138 138 138 138 138 Centered R2 0.358 0.520 0.541 0.557 0.568 0.557 0.502 0.243 0.229 0.537 Note: Coefficients that are statistically significant (at percent level) are shaded. The pvalues between brackets. Suffix 5 indicates five-year average. 34 Table 12. Explaining mean-normalized decile incomes (1988,1993, 1998): Adding regional dummies and inflation (GMM/IV estimation; dependent variable: decile mean income/overall mean income) First Second Third Fourth Fifth Sixth Seventh Eighth Ninth Tenth Openness5 -0.1306 -0.1662 -0.1712 -0.1741 -0.1571 -0.1265 -0.0811 0.001056 0.1772 0.8907 (0) (0) (0) (0) (0) (0) (0.006) (0.972) (0) (0) Expgdp5 0.1619 0.1764 0.1330 0.1092 0.0795 0.0233 -0.0360 -0.0922 -0.1174 -0.4849 (0.009) (0.004) (0.021) (0.056) (0.146) (0.666) (0.484) (0.061) (0.092) (0.201) Mean income -0.000009 -0.000010 -0.000012 -0.000013 -0.000012 -0.000011 -0.000008 -0.000001 0.00001 0.00006 (0.048) (0.030) (0.007) (0.001) (0) (0.001) (0.002) (0.593) (0.005) (0.007) Openness5*mean income 0.000017 0.0000216 0.000021 0.000021 0.000018 0.000014 0.000008 -0.000005 -0.00002 -0.0001 (0.004) (0.001) (0.001) (0) (0) (0.001) (0.024) (0.710) (0.004) (0.002) M2gdp5 0.0309 0.0390 0.0396 0.0038 0.0396 0.0441 0.0469 0.0471 0.0803 -0.4445 (0.378) (0.229) (0.200) (0.204) (0.161) (0.101) (0.07) (0.113) (0.132) (0.025)

DFI5 -0.0021 -0.0047 -0.0065 -0.0060 -0.0061 -0.0068 -0.0067 -0.0049 -0.0011 0.0435 (0.637) (0.213) (0.053) (0.094) (0.101) (0.100) (0.147) (0.318) (0.838) (0.110) DFI5*mean income -0.000002 -0.000001 -0.000001 -0.000001 -0.000001 -0.0000004 -0.0000004 0.0000003 0.00000001 0.0000083 (0.101) (0.095) (0.147) (0.088) (0.094) (0.237) (0.573) (0.985) (0.185) (0.195) Democracy5 0.0018 0.0076 0.0085 0.0088 0.0085 0.0081 0.0066 0.0042 -0.0036 -0.0498 (0.558) (0.003) (0.001) (0.001) (0.001) (0.002) (0.012) (0.146) (0.432) (0.010) Rint5 -0.0007 -0.0009 -0.0009 -0.0009 -0.0009 -0.0009 -0.0008 -0.0006 0.0021 0.0050 (0.013) (0.002) (0.004) (0.002) (0.002) (0.002) (0.013) (0.062) (0) (0.009) Linf5 -0.0193 (0.006) -0.0285 (0) -0.0300 (0) -0.0311 (0) -0.0292 (0) -0.0251 (0) -0.0171 (0.005) -0.0074 (0.254) 0.0475 (0) 0.1515 (0) Africa -0.0053 (0.890) -0.0323 (0.409) -0.0635 (0.081) -0.0850 (0.018) -0.0989 (0.004) -0.1052 (0.002) -0.1024 (0.003) -0.0612 (0.119)

0.0442 (0.423) 0.5255 (0.045) Asia 0.0622 (0.075) 0.0032 (0.927) -0.0342 (0.258) -0.0576 (0.050) -0.0760 (0.005) -0.0966 (0) -0.1146 (0) -0.1110 (0) -0.0493 (0.180) 0.4141 (0.036) Latin America -0.0733 (0.026) -0.1033 (0.001) -0.1229 (0) -0.1327 (0) -0.1378 (0) -0.1417 (0) -0.1357 (0) -0.0909 (0) 0.0450 (0.226) 0.8832 (0) 35 EEurope 0.1315

(0) 0.1343 (0) 0.1271 (0) 0.1122 (0) 0.0885 (0) 0.0592 (0.006) 0.0193 (0.313) -0.0291 (0.184) -0.1270 (0.001) -0.5358 (0.001) Constant 02924 (0) 0.4519 (0) 0.5936 (0) 0.7193 (0) 0.833 (0) 0.9607 (0) 1.1019 (0) 1.2608 (0) 1.3581 (0) 2.3871 (0) Hansen J 3.03 4.045 2.057 1.002 0.306 0.082 0.001 0.011 0.181 0.423 (0.0817) (0.044) (0.151) (0.316) (0.580) (0.774) (0.969) (0.917) (0.670) (0.515) No of obs 135 138 138 138 138 138 138 138 138 138 Centered R2 0.635 0.736 0.756 0.765 0.764 0.741 0.660 0.337 0.454 0.717 Note: Coefficients that are statistically significant (at 5 percent level) are shaded. Openness and government expenditure as share of GDP are

instrumented. GMM calculations performed using ivreg2.ado routine developed by Baum, Schaffer, and Stillman (2002). The p-values between brackets. Suffix 5 indicates five-year average. 36 Table 13: Explaining mean-normalized decile incomes (1988,1993, 1998): Changing the definition of openness to Dollar-Kraay measure (GMM/IV estimation; dependent variable: decile mean income/overall mean income) First Second Third Fourth Fifth Sixth Seventh Eighth Ninth Tenth Open_D-K5 -0.1052 -0.1501 -0.1495 -0.1399 -0.1155 -0.0776 -0.0248 0.0559 0.2336 0.4732 (0.055) (0.014) (0.012) (0.017) (0.040) (0.159) (0.642) (0.282) (0.002) (0.216) Expgdp5 0.1497 0.1518 0.1050 0.0772 0.0482 -0.0070 -0.0628 -0.1127 -0.1260 -0.2356 (0.033) (0.024) (0.093) (0.216) (0.412) (0.902) (0.234) (0.024) (0.089) (0.553) Mean income -0.000006 -0.000009 -0.000010 -0.000010 -0.000009 -0.000007 -0.000005 0.000001 0.00002 0.00004 (0.227) (0.138) (0.068) (0.049) (0.039) (0.071) (0.154) (0.738) (0.010) (0.196) Open_D-K5* mean income 0.000014 0.000022 0.000021 0.000020 0.000016 0.000012 0.000005 -0.000005 -0.00003 -0.00008 (0.019) (0.002) (0.003) (0.003) (0.005) (0.031) (0.276) (0.339) (0.001) (0.047) M2gdp5 0.0134 0.0249 0.0220 0.0161 0.0147 0.0195 0.0261 0.0363 0.0889 -0.2717 (0.701) (0.472) (0.529) (0.640) (0.647) (0.512) (0.336) (0.188) (0.080) (0.202) DFI5 -0.0066 -0.0084 -0.0107 -0.0106 -0.0106 -0.0106 -0.0093 -0.0052 0.0034 0.0688 (0.173) (0.050) (0.004) (0.005) (0.005) (0.007) (0.031) (0.266) (0.581) (0.010) DFI5*mean income -0.0000017 -0.000002 -0.000001 -0.000001 -0.000001 -0.000001 -0.0000005 -0.0000002 0.000001 0.000007 (0.215) (0.192) (0.289) (0.206) (0.184) (0.280) (0.572) (0.867) (0.472) (0.224) Democracy5 0.0025 0.0074 0.0083 0.0091 0.0090 0.0086 0.0070 0.0041 -0.0049 -0.0509 (0.417) (0.007) (0.003) (0.001) (0.001) (0.002) (0.009) (0.163) (0.285) (0.011) Rint5 -0.0005 -0.0007 -0.0006 -0.0007 -0.0007 -0.0007 -0.0007 -0.0006 0.0019 0.0034 (0.065) (0.017) (0.036) (0.028) (0.021) (0.018) (0.042) (0.075) (0) (0.055) Lninf5 -0.0170 -0.0254 -0.0261 -0.0266 -0.0244 -0.0202 -0.0122 -0.0034 0.0501 0.1104 (0.025) (0.001) (0.001) (0) (0.001) (0.004) (0.082) (0.643) (0) (0.016) Africa -0.0323 -0.0780 -0.1106 -0.1279 -0.1328 -0.1274 -0.1102 -0.0492 0.0976 0.6734 (0.468) (0.097) (0.013) (0.004) (0.002) (0.001) (0.003) (0.193) (0.073) (0.024) Asia 0.0566 -0.0053 -0.0434 -0.0684 -0.0870 -0.1050 -0.1176 -0.1014 -0.0068 0.4702 (0.177) (0.901) (0.263) (0.070) (0.011) (0.001) (0) (0) (0.866) (0.050) 37 Latin America -0.0879 -0.1260 -0.1498 -0.1591 -0.1605 -0.1586 -0.1447 -0.0892 0.0682 0.9997 (0.009) (0) (0) (0) (0) (0) (0) (0) (0.056) (0) EEurope 0.1194 0.1153 0.1046 0.0902 0.0690 0.0440 0.0099 -0.0287 -0.1054 -0.4324 (0.001) (0) (0.001) (0.001) (0.005) (0.033) (0.583) (0.150) (0.003) (0.005) Constant 0.2693 0.4381 0.5787 0.6967 0.8070 0.9307 1.0694 1.2348 1.3448 2.6328

(0) (0) (0) (0) (0) (0) (0) (0) (0) (0) Hansen J 0.291 (0.589) 0.068 (0.794) 0.165 (0.684) 0.473 (0.491) 0.758 (0.383) 0.840 (0.359) 0.721 (0.395) 0.249 (0.617) 0.068 (0.794) 0.325 (0.568) No of obs 138 138 138 138 138 138 138 138 138 138 Centered R2 (0.602) (0.710) (0.727) (0.734) (0.736) (0.717) (0.645) (0.343) (0.447) (0.692) Note: Coefficients that are statistically significant (at 5 percent level) are shaded. Openness and government expenditure as share of GDP are instrumented. GMM calculations performed using ivreg2.ado routine developed by Baum, Schaffer, and Stillman (2002). The p-values between brackets. Suffix 5 indicates five-year average. 38 6. Conclusions The effects of globalization on income distribution within rich and poor countries are a matter of controversy. While international trade theory in its most abstract formulation implies that increased trade and foreign investment should make income distribution more equal in poor countries and less equal in rich countries, finding these effects has proved elusive. Here we have tried to discern the effects of globalization by using the data from household surveys and by looking at the impact of openness (trade as the share of GDP) and direct foreign investment (as percent of the countrys GDP), on relative income shares of low and high deciles. We are thus able to chart as it were the effect of globalization on the entire income distribution in

both poor, middle-income and rich countries. We find rather robust evidence that at very low income levels, it is the rich who benefit from openness. As income level rises, that is for countries with survey-incomes of between $4-7,000 at international prices, the situation changes and the relative income of the poor and the middle class rises compared to the rich (top two deciles). It seems that openness makes income distribution worse before making it better or differently that the effect of openness on countrys income distribution depends on initial income level. These results run counter to simple Heckscher-Ohlin model with two types of labor. They are however consistent with a view propounded by Wood (1994) and more recently Kremer and Boskin (2003) that, with three types of labor (no education, basic, and highly skilled), openness in very poor countries might increase inequality by helping those with basic education, and leaving even further behind those with no education. Only when the poor become reasonably skilled, can the low deciles shares begin to benefit from increased labor demand; then inequality falls. This seems to provide an argument for free basic education as well as for the strong externality effects of more educated population. As for the other variables, we do not find any effect of direct foreign investments on income distribution. We find that democracy raises income shares of the middle deciles and leaves those of the top and the bottom unchanged (thus possibly explaining why synthetic 39 measures of inequality like the Gini coefficient have generally failed to detect an effect of democracy on inequality). Government expenditures and financial depth, as expected, do help increase income shares of the bottom and middle income groups and lower those of the top. Real rate of interesta topic which has surprisingly attracted very little attention and inflation are always pro-rich. Even middle classes lose (in relative share) when real interest and rate of inflation are high. In conclusion, increased trade seems to result in greater inequality, that is reduced income share of the poorest deciles in poor countries. Those who, according to economic theory and according to the policy prescription of international organizations, should benefit the most from increased trade appear, on the contrary, to be the losers in relative terms. The case for trade as an

engine of growth for the poorest of the poor is not however irretrievably lost. It must be based on trades impact on average incomes which, if sufficient, might lift real incomes of the poor as well. The case however cannot be made, we think, on trades favorable impact on income distribution. 40 REFERENCES Arbache, Jorge Saba (1999), How do economic reforms affect the dispersion and structure of wages? The case of an industrialising country labor market, paper presented at the 1999 Rayal Economic Society conference in England, typescript. Baldwin, Richard E. and Philippe Martin (1999), Two Waves of Globalization: Superficial Similarities, Fundamental Differences, January 1999,NBER Working Paper No. 6904. Barro, Robert (2000), Inequality and growth in a panel of countries, Journal of Economic Growth, vol. 5, pp. 5-32. Baum, C. F., Schaffer, M. E., Stillman, S., (2002), Instrumental variables and GMM: Estimation and testing, Working paper no. 545, Boston College Department of Economics. Available at http://fmwww.bc.edu/ec-p/WP545.pdf. Beck, T., G. Clarke, A. Groff, P. Keefer, and P. Walsh (2000), New tools and new tests in comparative political economy: The Database of Political Institutions, World Bank Policy Research Working Paper No. 2283, February 2000. Behrman, Jere, Nancy Birdsall and Miguel Szekely (2003),Economic policy and wage differentials in Latin America, Center for Global Developement Workling Paper No. 29, April. Available at http://www.cgdev.org/Publications/?PubID=29. Benarroch, Michael and James D. Gaisford (1997), Economies of Scale, International Capital Mobility, and North-South Inequality, Review of International Economics, vol. 5, No.3, pp. 412-428. Beyer, Harald, Patricio Rojas and Rodrigo Vergara (1999), Trade liberalization and wage inequality, Journal of Development Economics, vol. 59, pp. 103-123. Birdsall, Nancy and Juan Luis Londono (1997), Asset Inequality Matter: An Assessment of the World Banks Approach to Poverty Reduiction, American Economic Review, vol. 87, May, pp. 32-37. Birdsall, Nancy and Juan Luis Londono (1998), No Trade-off: Efficient Growth via More

Equal Human Capital Accumulation in Nancy Birdsall, Carol Graham and Richard Sobot (eds.) Beyond Tradeoffs: Market Reforms and Equitable Growth in Latin America, Washington, D.C.: Inter-American Development Bank and Brookings Institutions, pp. 111-145. Birdsall, Nancy and Amar Hamoudi (2002), Commodity Dependence, Trade and Growth: When Openness is not Enough, Center for Gloabl Development Working Paper No. 7, May. Available at http://www.cgdev.org/Publications/?PubID=7. 41 Bollen, K. and R.W. Jackman, 1985, Political democracy and the size distribution of income, American Sociological Review, 50, 438-457. Bordo, Michael D., Barry Eichengreen, Douglas A. Irwin (1999), Is Globalization Today Really Different than Globalization a Hundred Years Ago?, NBER Working Paper No. W7195, June 1999. Cornia, Andrea and Sampsa Kiiski (2002), Trends in Income Distribution in the post World War II period: Evidence and Interpretation, WIDER Working paper No. 2001/89, September 2001, downloadable from http://www.wider.unu.edu/publications/dps/dp2001-89.pdf. Craft, Nicholas (2000), Globalization And Growth In The Twentieth Century, Background paper to IMF, World Economic Outlook 2000, mimeo, January. Dluhosch, Barbara (1998), Globalization and European Labor Markets, October 1998, CEPR Discussion Paper No. 1992. Dollar, David and Aart Kraay (2000), Growth is Good for the Poor, mimeo, March 2000. Dollar, David and Aart Kraay (2001), Trade, Growth, and Poverty, Policy Research Worrking Paper No. 2615, Washington: World Bank. Dollar, David and Aart Kraay (2002), Growth is Good for the Poor, Journal of Economic Growth, vol. 7, pp. 195-225. Easterly, Bill, Roumeen Islam and Joseph Stiglitz (2001), Volatility nad Macroeconomic Paradigms for Rich and Poor Countries, in Jacques Dreze (editor), Advances in Macroeconomic Theory, New York: Palgrave. Fields, Gary (2001), Distribution and development: A new look at the developing world, Russell Sage Foundation and MIT Press. Freeman, Richard B. (1995), Are your wages set in Beijing? , Journal of Economic Perspectives, vol. 9 (Summer), No. 3, pp. 15-32. Galbraith, James K. and Hyunsub Kum (2002), Inequality and Economic Growth: Data Comparisons and Econometric Tests, mimeo, version April 5, 2002. Gradstein, Mark, Branko Milanovic and Yvonne Ying (2001), "Democracy and Income

Inequality: An Empirical Analysis", World Bank Policy Research Working Papers Series No. 2561, February. 42 Gradstein, Mark and Branko Milanovic (2004), Does Libert=Egalit? A survey of the empirical evidence on the links between political democracy and income inequality, Journal of Economic Surveys, September. Gundlach, Erich and Peter Nunnenkamp (1996), Some Consequences of Globalization for Developing Countries, August 1996, Kiel Working Paper 756, Institut fur Weltwirtschaft. Harrison, Ann and Gordon Hanson (1999), Who gains from trade reform? Some remaining puzzles, Journal of Development Economics, vol. 59, pp. 125-154. Higgins, Matthew and Jeffrey Williamson (1999), Explaining inequality the world round: Cohort size, Kuznets curve, and openness, June 1999. Available at www.ssrn.com. Kanbur, Ravi (1998), Income distribution and development, November, typescript. Kremer, Mark and Eric Maskin (2003), Globalization and Inequality, unpublished manuscript. Lejour, Arjan M. and Paul J.G. Tang (1999), The differential impact of the South on wage inequality in the North, CPB Netherlands Bureau for Economic Policy Analysis, The Hague, June 2000. Li, Hongyi, Lyn Squire and Heng-fu Zou (1998), Explaining international and intertemporal variations in income inequality, Economic Journal, vol 108, pp. 26-43. Lubker, Malte, Graham Smith and John weeks (2002), Growth and the Poor: A Comment on Dollar and Kraay, Journal of International Development, vol. 14, pp. 555-571. Lucas, Robert (1990), Why Doesnt Capital Flow from Rich to Poor Countries?, American Economic Review, vol. 80, No, 2, pp. 92-96. Lundberg, Mattias and Lyn Squire (1999), Growth and inequality: exctracting the lessons for policy-makers, mimeo, February. Lundberg, Mattias and Lyn Squire (2003), The simultaneous evolution of growth and inequality, Economic Journal, vol. 113, No. 487, April. Lustig, Nora and Ravi Kanbur (1999), Why is Inequality Back on the Agenda? Paper prepared for the Annual Bank Conference on Development Economics, World Bank, Washington, D. C., April 28-30, 1999. Available at www.worldbank.org/poverty/wdrpoverty/inequality.htm. Melchior, Arne, Kjetil Telle and Henrik Wiig (2000), Globalisation and Inequality: World Income Distribution and Living Standards, 1960-1998, Royal Norwegian Ministry of Foreign Affairs, Studies on Foreign Policy Issues, Report 6B:2000, October.

43 Milanovic, Branko (1994), "Determinants of Cross-country Income Inequality: An Augmented Kuznets' Hypothesis", World Bank PRE Working Papers Series, No. 1246, January 1994. Milanovic, Branko (2000), The median voter hypothesis, income inequality and income redistribution: An empirical test with the required data, European Journal of Political Economy, vol. 16, No.3, September 2000, pp. 367-410. Milanovic, Branko and Shlomo Yitzhaki (2002), "Decomposing World Income Distribution: Does the World Have a Middle Class?", Review of Income and Wealth, Vol. 48, No. 2, June, pp. 155-178. Milanovic, Branko (2002), True world income distribution, 1988 and 1993: First calculation based on household surveys alone, Economic Journal., vol. 112, No. 476, January, pp. 51-92. Milanovic, Branko (2004), Worlds Apart: Global and International Inequality 1950-2000, forthcoming. Ravallion, Martin (2001), Growth, inequality and poverty: looking beyond averages, World Development, vol. 29, n11, November, pp.1803-15 Robertson, Raymond (2000), Trade Liberalisation and Wage Inequality: Lessons from the Mexican Experience, World Economy, vol. 23, n6 (June 2000): pp. 827-49. Rodrik, Dani (2000), Comments on Trade, Growth, and Poverty by D. Dollar and A. Kraay, draft. Available at http://ksghome.harvard.edu/~.drodrik.academic.ksg/. Sala-i-Martin, Xavier (2002), The Disturbing Rise of World Income Inequality, NBER Working paper No. 8904, April. Available at www.nber.org. Schott, Peter K. (1999), One Size Fits All? Specialization, Trade and Income Inequality, October 31, 1999. Schultz, T. P. (1998). Inequality in the distribution of personal income in the world: how it is changing and why. Journal of Population Economics, vol. 11, No. 3, pp. 307-344. Slaughter, Matthew J. and Phillip Swagel (1997), The Effect of Globalization on Wages in the Advanced Economies, International Monetary Fund Staff Studies for the World Economic Outlook, December 1997. Spilimbergo, Antonio, Juan Luis Londono and Miguel Szekely (1999), Income distribution, factor endowment and trade openness, Journal of Development Economics, vol. 59, pp. 77-101. 44 Tang, Paul J.G. and Adrian Wood (1999), Globalisation, Co-Operation Costs and Wage Inequalities, January 1999, mimeo. Tavares, Jose and Romain Wacziarg (2001), How Democracy Affects Growth, European Economic Review, vol. 45, pp. 1341-1378. Williamson, Jeffrey G. (1996), Globalization and Inequality Then and Now: The Late

19th and Late 20th Centuries Compared, March 1996, NBER Working Paper 5491. Wood, Adrian (1994), North-South trade, employment and inequality: changing fortunes in a skill-driven world, Oxford: Clarendon Press. Wood, Adrian (1995), How trade hurt unskilled workers, Journal of Economic Perspectives, volume 9 (Summer), No 3, pp. 57-80. Wood, Adrian (1998), Globalisation and the rise in labor market inequalities, Economic Journal, September, vol. 108, pp. 1465-1482. Wood, Adrian (2000), Globalisation and wage inequality: a synthesis of three theories, August. Available at www.ssrn.com. Yusuf, Shahid (2003), Globalisation and the Challenge for Developing Countries, Journal of African Economies, vol.12, AERC supplement, pp. 35-72.

A final decision on the contours of community radio policy for India has been referred to a group of ministers.

Hoot Desk

In a cabinet meeting the UPA government today considered a proposal to approve phase II of the implementation of community radio in the country and decided to refer the matter to a Group Of Ministers. Jaipal Reddy, the Minister for Information and Broadcasting who will head the GOM said that hopefully it would arrive at a decision in a few weeks time. The ministers in the group will consist of Sharad Pawar, Minister of Agriculture, Dayanidhi Maran, Minister for Communications, Mr Reddy himself, and possibly the finance minister and a couple of others. Only the three mentioned have so far been named. In Phase II the plan is to allow both non governmental organisations as well as civil society groups to get licences to start radio stations. The matter on which there was a difference of opinion in the cabinet was on allowing such stations to have a revenue stream, and on the question of the range of broadcast to be permitted. Mr Reddy said that whereas he, Mr Pawar and the finance minister Mr Chidambaram were of the view that

some revenue stream should be allowed, others differed because these will not be paying licence fees like the FM stations. Mr Reddy said his proposal was for permitting five minutes of advertising per hour and some sponsored programming. He said the first phase of community radio had failed to take off because no revenue stream had been permitted. On the question of approvals, applications from NGOs and civil society bodies for licences will have to be decided within three months and will be referred to the ministry of home affairs for its comments. However, failing clearance before that, at the end of three months a committee of officials from all the concerned ministries headed by an official from information and broadcasting will take a final view. No ministry will be allowed to stall a proposal indefinitely, he said. Theoretically the number of frequencies available is huge, but a decision on the range to be permitted will have to be taken to know exactly how many frequencies will become available. Political news and current affairs will not be permitted, but informative broadcasts aimed at various community users will be. It would seem that the long wait for a viable policy on community radio, just got longer.

BLOGS If blogs are to be taken seriously, they should live up to the standards of accountability and reliability of the mainstream media that they so deplore.

Sajan Venniyoor

In June this year, the youth magazine JAM (Just Another Magazine) ran a rather unflattering story on the Indian Institute of Planning and Management (IIPM). IIPM, seen as a somewhat cocky challenger to the IIMs (Dare to dream beyond the IIMs!), and known for its lavish advertising budget, was not amused to read The Truth about IIPMs Tall Claims. The JAM story questioned IIPMs high ranking in the Outlook-Cfore B-school ratings of 2003 (which IIPM reportedly quotes in many of its ads). JAM magazine points out that the research agency, Cfore, had removed IIPM from its ranking as it had received "serious complaints about the veracity of information given by them." Furthermore, says JAM, its campuses were nothing like the vast and modern architectural wonders promised by its full-page ads, its faculty wasnt drawn from Harvard, Columbia, Yale, Insead as stated, its placement record wasnt all that hot, nor did its students get the prodigious packages promised by the institution. And - whats worse - its MBA degree was not even recognized in India.

The expose was picked up from JAMs website by blogger Gaurav Sabnis, an IIM Lucknow graduate. In language much less temperate than JAMs, Sabnis poured scorn on IIPM, accusing it of sc**wing around with peoples lives and careers. He cast doubts on the academic antecedents of its promoter, Arindam Chaudhuri, and - in case one didnt get the point - titled his posting The fraud that is IIPM. Things went rapidly downhill after that. Blogs sprang up overnight in defence of IIPM. Eschewing the niceties of debate, OracleCoder, Jassi, Raghuveer Srinivas, IIPMStudent9 and others savaged Rashmi Bansal, the editor and publisher of JAM, and Gaurav Bansal. Four letter words flew like confetti as they questioned Bansals credibility, credentials and sexual orientation. A shaken Bansal responded in her blog Youth Curry on 7 October, quoting at length from one of the anonymous bloggers who, after calling JAM magazine a dubious small-time publication, had gone on to accuse her of running a smear campaign against IIPM at the behest of a rival institution. "Aaj Tak," said the blog, "ran a sting operation where it was revealed on camera that JAMMAG accepted money in cash from [name withheld] group of educational institutions to write a negative story on IIPM" "There has been no story on Aajtak re JAM magazine," countered a furious Bansal, "There is no expose because there is nothing to expose." The blog war took a new turn when someone claiming to be the head of the legal and compliance counsel of IIPM slapped legal notices on Gaurav Sabnis and Rashmi Bansal, as well as some of their supporters and fellow bloggers. Sabnis woes did not end there. Sabnis, who worked for IBM, was informed by his bosses that IIPM was putting pressure on the company, to force him to pull his posts. He insists that IBM never directly asked him to remove his posts, but apparently the pressure got worse. His company, wrote Sabnis in his blog, began to get some bizarre threats from IIPM. The crunch came, says Sabnis, when the Dean of IIPM wrote to a senior executive at Lenovo (IBM) saying that "the IIPM Students Union had decided that if my blog posts were not deleted, then they would gather all the Thinkpads they had been given by the institute, and burn them in front of the IBM office in Delhi." On Oct 10, Gaurav Sabnis posted a cheerless update on his blog. "I have resigned from IBM," he wrote. He added a disclaimer that the decision was "entirely my own and I was not asked by IBM to resign, nor was I pressured in any manner" The face-off between blog and B-school had claimed its first victim. In spite of the threat of legal action, none of the bloggers have removed the offending posts and links from their blogs. All the same, there is an undercurrent of unease within the community. Many of them choose to see the incident as a threat to their freedom of speech. Heres a powerful institution trying to gag a small but determined group of Truth Seekers (as one of the bloggers calls himself/herself), while the mainstream media stands idly by. Indeed, there is much angst over the role - or absence thereof - of the MSM. The bloggers have an easy explanation for this, as articulated by one of them: "Well IIPM

is one of the leading media spenders... thanks to their full-page ads. So, its money talking all the way!" To be fair, the mainstream media - Hindustan Times, NDTV, Pioneer and Indian Express - did take note of the ongoing spat, with most of them sympathetic to the bloggers. But it says much about the navel-gazing quality of the virtual world that few bloggers realize that blogging is a minority interest, and that even in the wired West, its only when they bring down a Dan Rather that their online crusades merit serious media or public attention. JAM magazine is very much a part of the mainstream, and it broke the story with some degree of restraint, but that cannot be said of all the bloggers. Gaurav Sabnis admits that the language he used "may be a bit harsh, with words like "sc**w around" and "crap" which may hurt some people's sensibilities, but I think they were appropriate in the context." Well, IIPM doesnt think so. Nor, I suspect, would any mainstream journal. Thats where the bloggers defense of unbridled blogging falls apart. Article 19 is invoked with much passion by the community, but even Article 19 places reasonable restrictions on free speech. Gaurav Sabnis is a seasoned blogger, and a fairly moderate voice in the Indian blogosphere. But surely he cannot be so disingenuous as to assume that he can - in a public forum - arbitrarily question the bona fides of an institution and its promoter, in a manner that could potentially damage its business, and then claim the protection of the right to free speech. Legally, IIPM is well within its rights in taking the blogger to court, says an advocate quoted by the Express. The organisations reputation is tarnished by such statements and if it continues to be eroded, it can resort to civil and criminal procedures under defamation laws." Sabnis is critical of the anonymous blogs that sprang up overnight in defense of IIPM, but thinks nothing of bolstering his own case with links to the anonymous Truth Seeker (an IIPM ex-student as well as ex-employee), whose blog came up at much the same time. If blogs are to be taken seriously as an alternative medium, they should measure up to the standards of accountability and reliability of the mainstream media that the bloggers so deplore. Not so long ago, a fairly popular blog took pot shots at that media behemoth and everybodys favourite target, the Times of India. The blogs readers were much amused; the Times less so. A legal notice was duly slapped on the blogger - a perfectly valid one in this case. Discretion prevailed over valour, and the blog closed down voluntarily. There was some outcry over the strong-arm tactics of the Times, but what is significant is that no attempt was made either to defend or substantiate the hostile comments made in the blog. When the IT Act 2000 - which is technically a law, though not yet in effect - comes into force, bloggers will face bigger challenges than dubious legal notices from the rich and powerful whom they challenge at will today. Just because you are the underdog does not mean that you are always right.

Information law: dead on arrival?

As resistance gathers, the bureaucracy seem determined to destroy the heart of the Act and are now asking that file notings be kept out of its ambit.

Reprinted from the Indian Express, October 1, 2005

Maja Daruwala The fate of the new Right to Information Act, 2005, may turn out to be much like that of many a girl child in India damned at birth. Hardly has the law been born than the government who parented it is working hard to strangle its unwanted child. The Right to Information Act is a powerful tool. It enables citizens to get most information held by the government. The Act makes real the fundamental right to know. It recognises the difficulties associated with getting information out of the bureaucracy. So it lays down a large category of information which the government must put out without being asked; another category which the government must give if requested by a citizen; and finally a residuary category which wont be given unless it can be shown that it is in the public interest to give it. Very importantly, the Act lays down penalties for unjustifiably withholding information. Right at its birth, the president suggested that all documents and communications emanating from the Rashtrapati Bhavan should be exempted from disclosure. This did not happen. But taking encouragement from that early signal, other less exalted agencies are jockeying to be exempted. As resistance gathers, the bureaucracy seem determined to destroy the heart of the Act and are now asking that file notings be kept out of its ambit. Of course they will want this because here lies the key to government power: unchallenged discretion to make decisions without accountability.

That sacred creature, the government file, usually has two halves: on the right side is the correspondence and materials. This could be someones grievance, a project proposal, a tender, a concept for a policy, anything the government needs to consider. On the left is the closely held note sheet, which records how the proposition was examined. The advice given in the notings has to be strictly in accordance with laws, rules, norms, and orders. The concerned officer puts down views, advice and recommendations. The file then goes up the line to the decision-making authority. On the way the hierarchy adds more notes, putting their signatures against their views. With all advice in hand a justifiable decision must be made. Cumulatively, the note sheet reflects the mind of government, bares the intention of an individual officer and whether his advice and consent were grounded on established rules. File notings, then, are X-rays of government functioning. They are proof of fair play and reason , or dishonesty, bias, and negligence. They are the shield that most honest bureaucrats wish they had and the sword that dishonest ones fear. It is only by allowing thorough public scrutiny of the evidence of how the government works at every level that corruption can be fought. File notings fall squarely within the definition of information in the Act. The nature of file notings is inevitably that of advice/opinion, and these are explicitly covered. To twist the definition to exclude file notings would be to destroy the Acts legislative intent. Yet that is what is being sought. Despite clear enunciation that file notings fall within the Act, the Department of Personal and Trainings website, says that "information means any material in any form including records, documents...but does not include file notings". No doubt, as some public servants seek to wriggle out of the inconvenience of explaining their actions to the public, there will be self-servicing internal opinion given to the government which urge the withholding of file notings. If they succeed, people will once again be shut out from their own governance as has been the culture for the last 58 years. This must be resisted. Of course the government has the power to amend the law. But the Act also clarifies that amendments can only be made to remove any difficulties that may arise in giving effect to its provisions and to further the objectives of openness and transparency. Amendments cannot be used to defeat the very purpose of the law. During parliamentary debates the government promised to amend only in the light of experience. But even without waiting to garner experience, it is now eager to pacify powerful bureaucrat lobbies with retrograde amendments. In a democracy people, not governments, are supreme and this is the truth the governing classes cannot stomach; they cannot imagine their performance being scrutinised by the great unwashed. The Right to Information Act was fuelled by the energy of very poor and vulnerable people who created a movement out of their desperation for good governance. It was passed with the intent of creating accountability and putting power where it belongs in the hands of people. It was designed to reduce the discretions and bias that plague government decision making. To compromise these principles would be to dishonour the aspirations of the most vulnerable among us.

Stretching the border of taste

There ought to be a level playing field for the two principal arms of the media - print and electronic. If sexually explicit content is the issue, TV is certainly more guilty than newspapers.

Reprinted from the Pioneer, August 21, 05

Chandan Mitra That the Supreme Court has admitted a PIL against the unfettered publication of allegedly sexually explicit material by newspapers and sent notices to two leading dailies and news agencies probably reflects the widespread concern coupled with helplessness at the growing trend to cater to the prurient instincts of readers. It's early days yet and thus difficult to gauge the direction the matter will take. No doubt, the issue of Press freedom will come up in a major way in case there is any move to classify newspapers into categories in accordance with their allegedly explicit content as demanded by the PIL. The Supreme Court has also invited a submission by the Press Council of India on whether certain guidelines can be framed for this purpose. On the face of it, the idea of certifying newspapers into categories like 'Adults Only' or 'Parental Guidance Advised' sounds rather far-fetched if not outrightly hilarious. It will also not be easy to codify 'sexually explicit' unless a censor board official, empowered to use discretion, is made to sit in every newspaper office and scan through its contents on a daily basis. I believe the apex court will not entertain any such suggestion as the idea of

posting a moral policeman in newspaper offices is neither desirable nor practical and will invite serious charges of curbing media freedom - anathema to the functioning of a mature democracy. India cannot be subjected to China-type regimentation or Saudi Arabia-like fundamentalist repression. However, some issues arise straightaway from the PIL. At various seminars, I routinely face a volley of questions from the audience about the publication of semi-nude photographs and titillating articles, many of which undoubtedly transgress the borderline of what is commonly understood as good taste. I have always replied to these accusations with a simple observation: Nobody forces a reader to subscribe to a newspaper that offends his/her sensibilities. The paradox is that the circulation of the allegedly prurient publications seems to be steadily increasing while the more sober papers, including The Pioneer and The Indian Express, are not growing at the same pace. In other words, there is a strong element of hypocrisy about this concern over alleged prurience. There is, in my opinion, an even bigger problem with TV. Some of the programmes that are regularly aired by both national and foreign channels truly stretch the borderline of taste. I never watch Fashion TV, but the other day, while surfing, I did see it for a few minutes and was appalled by what I saw. Opinion was divided when Ms Sushma Swaraj, as I&B Minister in the NDA Government, banned this channel. After what I saw last weekend, I was certain that the ban should never have been lifted and, actually, ought to be reimposed. Arguably, the visuals did not promote sexual activity. But the procession of models flaunting their assets without even a fig leaf on their torsos cannot be justified on grounds of aesthetic license. Late night shows, many of which are often repeated during the afternoon when children are back from school are even worse. Some channels have taken not just to re-creating murders but also rapes! Unlike print, TV has no regulatory body overseeing its contents. True, the Press Council does not pre-judge contents of a newspaper and acts only when a complaint, usually of defamation or libel is registered, but an aggrieved citizen at least has the option of complaining about any printed news item. Somebody defamed by TV has no recourse whatsoever. This is a major lacuna in the regulatory framework, but even 15 years since cable TV mushroomed in India, the Government appears unable as well as unwilling to act. Occasionally, there has been talk of setting up a Media Council to cover both print and TV or a Broadcast Authority of India on the lines of the Advertising Council, but none has fructified. During the recent controversy over India TV's sting operations, the Government said it proposed to legislate to control such reportage. Nothing has come of that either. A fortnight ago, responding to a Rajya Sabha debate on obscene content on TV, the Government again shared MPs' concerns and assured steps to contain runaway prurience.

It is not my argument that TV should come under the ambit of censorship. A majority of channels do not transgress the borderline of good taste and, in fact, telecast informative, wholesome content. But there ought to be a level playing field for the two principal arms of the media - print and electronic. If sexually explicit content is the issue, TV is certainly more guilty than newspapers. Of course, here too, the same argument can be made. Just as readers have the choice of not subscribing to a newspaper they don't like, they also have a remote control gadget in their hands and are, therefore, not compelled to watch offending channels. Regardless of the view the Supreme Court eventually takes on the PIL in question, I believe there are certain issues the media itself has to handle. In his address at the 2nd Convocation of the Pioneer Media School last Monday, TV celebrity Rajdeep Sardesai made a few pertinent observations. The basic thrust of his query was: What constitutes news? He pointed to a series of news bulletins on most channels over the last weekend that focused exclusively on just two items - actress Karisma Kapoor's legal dispute with her estranged husband and selection of the Indian team for the forthcoming tri-series in Zimbabwe.

He rightly wondered if there was nothing else happening in the country for the channels to devoted almost the whole bulletin(s) to just these two relatively unimportant matters. No doubt people are interested in knowing the Indian cricket team's composition, but does that merit dedication of 30 minutes in bulletin after bulletin, to the exclusion of other news? And, is it of any material consequence to the people of India if Karisma Kapoor is not allowed to travel abroad with her four-month-old child? The steady dumbing down of news, both by print and TV, but especially TV, is what should seriously concern not just media professionals but also all thinking people in the country. Although not directly related to the PIL on explicit content, there is a connection between the two: They feed on each other. Because news is progressively becoming conditioned by entertainment, it is only one logical step further for newspapers and TV channels to promote titillating matter. There is a chicken and egg question here; it's difficult to pinpoint what came first. But I have no doubt that the media itself must contain the fatal attraction of dumbing down news if it is to ward off public anger and, eventually, judicial reprimand. Freedom of the Press does not mean license to print quasi-obscene or vulgar material with purely commercial interests in mind. In the West, where society is much more permissive than India, publications are clear about their target readership. Thus, there are

tabloids and magazines almost specialising in explicit content while the mainstream papers and journals studiously avoid descent into prurience. In India, however, it is the mainstream rather than niche publications that are guilty of promoting near-pornographic material. Laws in this regard are much more liberal in the West, which means that the issue cannot be resolved by legislation. It is for the media itself to ponder the consequences of its current policies.

Community radio revolution in Uttaranachal

Using amplifiers and stereo equipment, the groups narrowcast their programmes in their village communities in an attempt to build 'listening clubs'

Women's Feature Service

Ruchika Negi in Dehradun

It has the beginnings of a revolution. Five community groups in the villages of Garhwal and Kumaon, in the state of Uttaranchal, are helping people to access government schemes, fight corruption and discuss their everyday problems via the community radio. These five groups are working to create knowledge network societies in remote villages, where both newspapers and television have no reach.

The Community Radio Uttaranchal initiative was started in 2001 with the help of Dehradun-based NGO, Himalaya Trust. Over 20 volunteers, representing five different areas of Garhwal and Kumaon, were given basic training in setting up a community radio network by the NGO. The project is supported by the international media agency, Panos.

"I think radio is a very important medium of information for people in the hills, where villages are located at a great distance from each other," says Rajendra Negi, a volunteer and co-founder of Hevalvani community radio (CR) project. "Half the villages are situated on heights where nothing reaches on time. The only thing people want to listen to is the radio. They are keen to listen to their own programmes, their own issues, their voices articulated in their language. Community radio is a medium that is our own."

Hevalvani CR is located in Heval valley of Tehri Garhwal, at a distance of about 40 kilometers from the nearest city centre, Rishikesh. The CR core group comprises of six young and active members from in and around the neighbourhood of Khadi, who work closely with the women's self-help groups to disseminate information on issues related to health, self employment and education.

"Recently, in a neighbouring village, an old woman who tried to raise her voice against rampant corruption in the village was beaten up by the revenue officer on the behest of the village head. This has been a recurring phenomenon in the village, so we decided to do something about it," says Dharmiyan Rana of Hevalvani project. "We interviewed the woman, the people from the neighbouring villages and made a radio programme which we even submitted to the district magistrate, along with a written appeal for further investigation into the case. Now people will think twice before committing such wrongs."

Raibar CR, situated in Balganga valley, operates from two distinct areas, the town centre of Chamyala and Bangaon - a far off village located on a mountain ridge - which is about an hour's climb through the forest. The peculiar location of the radio group makes it possible for the members to collect relevant, topical information from the well connected, burgeoning township of Chamyala and disseminate it, with the help of their radio.

Mandakini ki Awaz (voice of Mandakini) CR is located in Mandakini valley, Pauri Garhwal. The members work from a small village, Bhanaj, about a five-hour drive from the nearest town centre of Rudraprayag. This group aims at creating an open and transparent administrative and governance information system and is thus collaborating with the panchayats (village councils) of the neighbouring three villages as well. The radio group mediates between the people and the governing bodies in these villages, in order to make policies, schemes and financial budgets, open for public discussion and scrutiny.

"Initially, when I heard about Mandakini ki Awaz, I was very sceptical about the initiative. What do we need another radio network for?" says Lakhpat Pundeer, a resident of Bhanaj, who recently joined the radio group. "But then I attended one session, organised by the radio volunteers with the panchayats, I understood its benefits. Here the villagers discussed the financial budget of the village: how much has been allocated for which project, who are the beneficiaries etc. Never before had these details been divulged to people directly. CR has been able to bring some level of accountability in the administration."

The CR groups strive to work not just towards community issues and concerns but also collect and document the oral history and traditions of the valley. This information is then strung into programmes which are aired.

"Although there is an influx of different kinds of information mediums, such as newspapers, television etc that apprise us about the world, there is no way that I can find out more about things that are closer to my reality. I am not interested in what is brewing in America! I may want to know more about employment opportunities, farming techniques, where to seek medical aid et al. Where can I get this information from? Community radio is my way of talking about the day to day needs, the issues and concerns of my people," says Archana Raturi, a young member of Hevalvani CR.

Using amplifiers and stereo equipment, the groups narrowcast their programmes in their village communities in an attempt to build 'listening clubs' where people can congregate to hear community radio programmes, debate and discuss issues. Gatherings like these facilitate a dialogue and initiate a process of questioning and unveiling. "Once we made a programme on panchayati elections. While narrowcasting, people suggested that rather than just focus on big issues why do we not also include simple, basic facts about voting, like how to cast one's vote, what is the colour of the ballot, what are the procedures that one has to follow?" recalls Vipin Joshi from Pradeep CR.

"The women, in particular, were keen to know all this because they are not very politically active. So, we remade our programme and that year the turnout of women voters was astounding. People bought our tapes even after elections were over and continue to play them in the villages," says Joshi.

Shristhi CR has its roots in Uttarkashi. Comparatively a younger group as opposed to the others, it is struggling to expand its voluntary support base and the members are trying to build support networks with the university students and other youth groups in the area. "We are trying to focus on issues of self employment, particularly among the youth. There are people from amidst us, who have chosen not to migrate to the cities and work on new, innovative enterprises here like sericulture, mushroom and vegetable cultivation etc. We make learning radio modules with their help, which we then play back at university, schools and in youth gatherings and clubs," says Sumitra of Shristhi CR.

For the volunteers, doing radio, thus, is not just restricted to serving the informational needs of their community. "Community Radio, in its true sense, can never flourish and grow in isolation. It is a people's movement and thus my role is not just restricted to providing information. It is a participatory process where users of knowledge networks must also actively contribute to that flow," says Negi.

Ham radio to the rescue in tsunami-hit Andaman

"The phone links had disappeared, so I started using my radio set to connect with people in mainland (India) and giving information about people in Port Blair."

Hindol Sengupta

Indo-Asian News Service

Port Blair, Jan 2 (IANS) When tsunami waves broke all communication lines across India's Andaman and Nicobar Islands, ham radio came to the rescue.

Bharati Prasad, India's foremost amateur ham radio operator, was attempting a new transmission record here when the Dec 26 earthquake followed by the giant waves hit the archipelago.

Faced with a massive humanitarian crisis - with hundreds dying and islands with thousands of people completely out of reach - Prasad quit her project and jumped into action.

"In situations like this, the only thing that works is radio," said Prasad, one of the 20,000 Indian ham radio operators - people who run powerful radio transmission sets connecting with fellow enthusiasts across the world.

"The phone links had disappeared, so I started using my radio set to connect with people in mainland (India) and giving information about people in Port Blair."In fact, I was one of the first people to get in touch with a radio guy in Thailand who told me what was happening there barely minutes after the waves hit (the beach town of) Phuket."

More than 700 people are confirmed dead and hundreds still missing in the Andamans a week after the tsunami ravaged coastlines across Asia, killing more than 120,000.

Sitting in the room in the Port Blair hotel - two cracks near the door bear ugly testimony of the quake - Prasad spoke about how the first people who used her facility were bellboys and bearers at the hotel.

"There was nothing else. There was some talk of satellite phones but that was way beyond their reach. This was free," smiled Prasad, 46, looking like a mother who is a little harried, in her brown cotton sari and thinning black hair tied in a small bun.

"They brought me slips of paper which had the names and the phone numbers of their family members and I asked by radio contacts in Chennai, Delhi and Mumbai to make phone calls and tell them that their boy is safe," she said.

So impressed were local administrators by her work that she has been asked to set up eight radio centres throughout the region, including remote, and worst-affected, islands like Car Nicobar and Campbell Bay.

"The radios are really helping us," said Andaman police chief S. B. Deol. "There are providing invaluable information about conditions around the islands - who needs what, where. "Now within a day or two 15 radio operators (from the National Institute of Amateur Radio in Hyderabad) are coming here and we would be dispatching them to the remote islands. They will be our connectivity points."

Prasad, who goes by the codename VU2RBI in the three million worldwide radio fraternity, has been an enthusiast for more than 25 years. "I am a science graduate who got interested in ham radio because my sister and brother-in-law used to do it," said Prasad, whose 15-year-old son is one of the world's youngest radio operators.

"I got my licence after a couple of years and, since then, this has been the primary thing in my life. It wasn't easy getting the licence; you have to clear electronics tests and then get home ministry security clearance. I was one of the first women in India to get a licence."

This time, Prasad had come to Port Blair - the first person in 17 years to get permission to do radio work here - to get a record of contacting more than 35,000 people around the world through her radio set. "We wanted to know how connected radio could be even at a (remote) place like the Andamans."

But all that changed with the tsunami. "Obviously disaster relief first," said Prasad. "In fact, I believe that the ham radio was invented to help in crises like this when nothing else works."

FMs hour of glory

FM radio rose to the Mumbai floods as no other media could. Will local relevance, FMs strength, be the first casualty of the proposed FM expansion?

Sajan Venniyoor "It was FM all the way as far as media was concerned," wrote Harini Calamur of Mumbai at the height of the floods in her city. "I really dont know anyone who watched too much TV. Around 70% of Mumbai barely had electricity. Most cable guys were down." Harini, a TV producer and filmmaker, was responding to a posting by self-confessed brain-dead techie Sameer Gharat in his blog, The Opti-Mystic (http://www.optimystic.net/blog/). On 28 July, Sameer had written, "The cynics that the [TV] news channels are, all they could concentrate on was the failure of the administrative machinery. The worst of the lot was Star News. It was constantly beaming stale pictures of waterlogged roads and announcing the crashing of Mumbais Shanghai dreams. All it concentrated was the question, "Why?!" rather than asking themselves "How" they could help make an actual difference on the street level." Star News, the Mumbai based TV news channel, reportedly had a team of 30 reporters and four OB vans deployed across the city, and claims to have done "far more justice to the havoc than any other channel." Much of this was probably wasted on Mumbaikars, as most of the city went without electricity and cable for days on end. "The SMS services offered by these news channels are a mere farce!" wrote the OptiMystic. "With electricity and cell-phone networks down or erratic, tickers scrolling at the bottom of the TV screens are the last thing people stuck on the roads are going to refer to! These are merely ways to generate revenue since each SMS is charged at a premium rate rather than the usual Re.1 charge. At the same time, the FM channels served us well by disseminating essential and latest information (area-wise flood warnings, traffic

situations, SOS calls from listeners, etc.) on a continuous basis. A tip of the hat to the FM channels and rotten eggs on the faces of the news channels!" For all those who had sneered at commercial FMs "corporate-controlled, cookie-cutter, bland, mass-market garbage", the response of FM channels to the floods in Mumbai was an eye-opener. The Economic Times said, "FM radio proved its worth when everything else seemed to be drowning." ET recorded several instances of the impact of FM radio. "A Honda City driver stuck in the swirling traffic rang in to 'Radio City' on Tuesday to give his location and car number to offer his car phone charger to those whose cell phone batteries had died out. Dozens of commuters listening in made their way to the Good Samaritans car to bring their handsets back to life!" (Tuning in, to the people - 29 July 2005). The paper reported that most radio stations had suspended or reduced their music content to turn themselves into "information points, help-lines and morale boosters all rolled into one." Of the four private FM channels in Mumbai (Radio Mirchi, Radio City, Red FM and Go FM), Go-92.5 was undoubtedly the channel of the moment. Says Bandana Mukhopadhyay, Mumbai resident and former AIR Director, "Go FM was the first off the mark, quickly changing its programmes to provide useful information to the public. The other channels soon followed, but Go FM did an absolutely superb job of giving local information." The ET report went on to say, "On Wednesday, when people were still stuck in the deluge, an RJ with FM station 'Go 92.5', cruised around the western suburbs, giving a running commentary on which roads and localities were still under water and how to avoid snarled junctions. Another RJ on the station kept reminding listeners in the cozy environs of their homes to help those in distress outside in the streets." (ET comes from the same stable that owns Radio Mirchi). In December 2004, when the tsunami struck the Andaman & Nicobar Islands, radio played a similar role in the archipelago. "Battery-operated radio sets are possibly the only source of information and entertainment for the people in the Andaman and Nicobar Islands who are trying to get on with their lives after the tsunami disaster," reported Satarupa Bhattacharjya in The Pioneer (When radio kept them connected - 7 Jan 2005). "Since electricity supply, disrupted by the killer waves, was restored only by Thursday [6 January] on the island chain, the affected people have not been able to watch TV for over a week." The same tale was repeated across the tsunami hit areas of the Indian Ocean, with thousands of portable radio sets being distributed in Banda Aceh (Indonesia), coastal Sri Lanka and the Maldives. Indiantelevision.com (7 January 2005) reported how Commercial Radio Australia shipped 50,000 AM/FM radios and six transmitters to the affected areas. It was weeks before electricity supply was restored, and TV screens flickered back to life. Consequences of deregulation

When disaster strikes, it is not often that corporate radio rises to the challenge. Radio monoliths and the demands of the bottom-line have destroyed much original and local programming on radio channels across the world. Why did Go 92.5 do a better job in Mumbai than the other channels? Surely it was not because Go had more money to throw around. (As far back as June 2003, after having suffered a reported loss of Rs.110 million in the preceding financial year, Go FM had threatened to close down and even issued a conditional notice to government). Perhaps it had something to do with the stations degree of local commitment. Writing on the sweeping changes wrought by de-regulation of radio in the United States (Consumers Trampled As Telecom Industry Runs Amok, Coastal Post, 1 Aug 2005), Sandy Leon Vest, a radio journalist with public and community radio, says, "Study after study has documented that profit-driven media conglomerates are investing less in news and information and that local news in particular is failing to provide listeners and viewers with the information they need [.]" She writes that the "trend of corporate modeling has resulted in the near extinction of true community radio and the erosion of local programming, staff morale and political integrity. Corporate allegiance to the bottom line has forced local TV and radio stations, often owned and/or underwritten by non-local corporations, to produce less local public affairs programming and to hire less local staff." Go FM is run by Mumbais own afternoon paper, Mid-Day. Big corporate media houses run the other FM channels in Mumbai, with sister operations in several other cities. Red FM, from India Todays radio wing, runs stations in Mumbai, Delhi and Kolkata. Radio City has channels in four cities, and Radio Mirchi (from the Times of Indias Entertainment Network) has 7 channels across India. Between them, the Big Three run 66% of all the private FM channels in the country. The erosion of local programming observed by Sandy Vest is inevitable on the channels run by these media conglomerates. The only private FM channel in Bangalore, Radio City, is a Bollywood channel, broadcasting Hindi film hits in a south Indian city; and youll have to listen long and hard to hear a Marathi programme on Radio Mirchi Pune. In spite of the ban on news, and the resultant public hand-wringing and breast-beating by private broadcasters, theres plenty of news on private FM. Tune in to City Scan on Go 92.5 FM. As the Go website (http://www.go925fm.com/scan.asp) says, the channel gives you all the news you can use. "Chris [RJ Christina Fernandes] gives you a round up of events, sports and stocks through the day! Get the pulse of Mumbai city delivered to you hot off the press on the hour every hour!" The fact is, all private FM channels across the country carry local news when it suits them. And the ban on hard (read political) news, as Bandana Mukhopadhyay points out, may not be such a bad thing after all. "Theres a fine line between news and information. News, as a rule, is easy to do on radio, compared to broadcasting local information. News could quickly come to mean national and international news-feeds from Reuters and PTI, which is not what communities need."

Phase-II of FM licensing, notified recently, promises to put 336 channels in 90 cities. In this affordable revenue-share regime, it seems more than likely that Mid-Day, like the other players, will try to extend its network across the country, snapping up frequencies in many cities. (Mid-Day certainly has - or had - national ambitions. Mid-Day Multimedia Ltd. had won licenses in three cities during the first phase of FM licensing in 1999. Due to high license fees, they were unable to start their operations in Delhi and Chennai). It is too early to speculate on how this will affect their commitment to local programming or allegiance to the bottom-line, but if the de-regulation of radio in the United States in 1996 is any indication, local relevance could be the first casualty of FM expansion. Following de-regulation in the US, writes Sandy Vest, "a mere ten companies began to dominate two-thirds of the radio audience with two companies, Clear Channel Communications and Viacom (owner of Infinity Broadcasting), controlling 42 percent of listeners and 45 percent of radio industry revenues. Radio monoliths such as Clear Channel have driven out minority radio station owners and made it difficult for noncorporate artists to get airtime on commercial radio." The most notorious example of the perils of excessive concentration of radio ownership comes from Minot, North Dakota, where Clear Channel owns six of eight commercial stations. During a chemical spill in January 2002, emergency workers were unable to reach a live person at any of the stations. Clear Channel owns 1200 of the 10,600 commercial radio stations in the US. In 2002, the Future of Music Coalition had noted how "all radio markets are dominated by four radio companies controlling at least 70 percent of the radio audience - with concentration even greater in smaller markets." Even with just 21 private FM channels in India at present, the situation is depressingly familiar; it is likely to get worse with 336 channels in 90 cities. Barbara Skerath, former Director of Deutsche Welle Radio Training Centre and longtime observer of Indian media says, "330 channels to be licensed in 90 cities - to me, that sounds more like mass instead of class. As far as I know, private radios in India have been limited to - or have limited themselves to - mere entertainment, glamour gossip and music. So whats the difference between one or ten or 330 stations which all offer Hindi Pop or film music? If you have heard one channel, you have heard them all." As Harini Calamur writes in her blog (Radio Go Go, http://calamur.org/gargi/), "FM radio is a local medium. Serving a local population. [] Unfortunately, the localisation of content and advertising on radio is still a long way away. All channels are me-too clones. With RJs who speak as though they picked up an accent while passing by the airport. With music that is much the same." Skerath points out, "In an environment where the media are money driven, any nonsense will go on air as long as it catches more attention than the other competitors. We have seen such a situation in Berlin where five or six private [radio] stations were competing for ratings and commercials. Programmes soon gave way to all kinds of absurd, shocking

or obscene games and talk shows just to gain - albeit short-lived - attention. For example, listeners were told to demolish their bathrooms while on air and the one who did it most effectively or most dramatically gained a ticket to some holiday destination. Hundreds ended up with destroyed bathrooms and nothing in return" Radio in India hasnt plumbed those depths yet. But the 336 new channels (which will more than double the number of public and private radio stations we have in India at present) are dramatically going to change the radio landscape in the country - though not necessarily for the better. As Brecht said in the very early days of radio broadcasting, and as the Mumbai floods have proved now, "Radio could be the most wonderful public communication system imaginable, a gigantic system of channels - could be, that is, if it were capable not only of transmitting but of receiving, of making listeners hear but also speak, not of isolating them but of connecting them."

Project: INFORMATION TECHNOLOGY AND RURAL EXTENSION IN INDIA Collaborators: Tamil Nadu University of Veterinary and Animal Sciences Chennai, Tamil Nadu, India College of Agriculture and Life Sciences Cornell University Ithaca, New York USA Current status (2002): Telecenters have been opened in several locations. Community surveys focusing on information needs have been undertaken. PROJECT BACKGROUND AND JUSTIFICATION The focus of this project is on developing the software infrastructure to maximize the impact of a networked group of rural information centres in south India. The Tamil Nadu Veterinary and Animal Sciences University (TANUVAS) has a three part mission: (1) it conducts academic programmes for baccalaureate and post-graduate degree students; (2) it conducts research related to animal sciences and related areas such as food production and rural entrepreneurship; and (3) it carries out extension

programmes for the benefit of rural dwellers, small businesses, cooperatives, various nongovernmental organizations, and government officials. This proposal is particularly related to the third part of the Universitys mission listed above: TANUVAS has a mandate to reach out to rural areas of Tamil Nadu to help communities and families improve their social and economic well-being. A substantial part of this extension effort is devoted to information and training programmes related to milk production, rabbitry, sheep and goat husbandry, small scale fisheries. One of the principal objectives of these efforts is to enable rural people to become small scale entrepreneurs and to insure that they have expert information support to help them sustain their enterprises. For example, the TANUVAS extension programme trains self-help rural womens groups in dairying, fisheries and other small-scale ventures, and provides information support to these groups through their group-elected representatives. Networking problem Currently TANUVAS tries to maintain contact with rural farmers and entrepreneurs through 14 University Training and Research Centres (UTRC), 3 Farmer Training Centres, and 2 Krishi Vigyan Kendras (KVK-Farm Science Centres) all scattered in rural towns throughout Tamil Nadu. These outreach services are staffed with animal and agricultural specialists, most of whom hold post graduate degrees and professional level rank in TANUVAS. The primary means of contact with rural constituents at the centers is through face-to-face contact, either individually or in training programmes and workshops. While strategically placed to serve rural communities, the breadth of information services is largely limited to animal production issues. With the present system based heavily on labor-intensive personal contact, the centres do not reach a significant proportion of the population. Through a World Bank supported Human Resources Development Project, TANUVAS will connect each of the rural centers to the Internet. The linkage of these centres via modern information technologies to the University and to other expert resources will enable the centres to address more comprehensively and more quickly the information and communication needs of rural families. The intent of this project is to transform and institutionalize a group of these centres as Rural Information and Communication Centres (RICC). The Centres would expand the resources available to their constituent communities in two significant ways: (1) by making available a range of information and communication technologies (ICT) to the public; and (2) by increasing the breadth of information and training resources available, thus including current market data, family health and nutrition information, literacy materials, and e-commerce facilities. A major challenge is to link these centres with experts at TANUVAS using digital networks and thereby making more efficient the seeking and diffusion of information. Because of the added advantage of exchanges in this process, the information and other resources at the centres will be appropriate and relevant to the rural communities.

There are important research dimensions in this project: 1. The first is to develop needs assessment tools specifically related to information, training, and communication technologies that can be used before start-up and continually thereafter so that the products and services of the centres are relevant to the variety of stakeholders in the communities. This means, for example, having a system that will find out the information and training needs of women and unemployed people who are often overlooked in development initiatives. The formulation, testing and refinement of these tools will provide models for expanding this project into more UTRCs and to telecentre-like projects elsewhere in the world. 2. The second research aspect is the development of methods and guidelines for training the community in the uses of the RICC, recognizing that the availability and accessability of such information centers does not automatically result in their use by the community and by community groups. For example, It will be a long, long time before all the obstacles to access are removed and everyone will voluntarily and spontaneously make use of the RICCs. It is important to recognize that while many in the community may themselves never use the centres directly, they may be able to benefit from them. Take, as an example, the more than a quarter of a million women who belong to self-help groups in Indias Tamil Nadu state. Telecenters could train representatives of these groups, who then would serve as liaisons between the groups and the RICCs. The liaison persons might also serve on an advisory committee for the centre, an important vehicle for community participation in a centres operations. The results of this research will have relevance also for the thousands of telecentres being established in Asia, Africa and Latin America. 3. The third area of research deals with methods for adding value to information resources so that they will be valuable and user-friendly for the surrounding communities. There are many data bases and other information resources in India and across the world that potentially have relevance for Indian communities. For example, there are more than 12,000 www files related to health, but most of them are in the English language. An important dimension of this project is the role of the University as an information provider and value-adding hub for a rural communication network. In this case, the University, in its fundamental role of public service to the society, will be the key institution that will adapt demand?driven information to suit the needs of the communities served, providing information products that are accessible to the communities through the network of RICCs. To illustrate with another example, Infectious Bronchial Trachitis (IBR) in cattle and small ruminants is prevalent in both India and the USA. To control this, a vaccine F5 is being used in dairy farms in rural New York. The scientists at TANUVAS are continuously trying to isolate the virus and produce vaccine for IBR disease. No vaccine has yet been produced in India to control the resulting mastitis problem. Originating with Cornell University scientists, knowledge related to this problem can be widely shared among veterinary technical people and ultimately farmers if it can be processed through the value adding hub at TANUVAS.

In this project we will test how universities can serve as support institutions that facilitate the introduction of Internet to rural and most disadvantaged areas. This project will experiment with how the Internet can be used and combined with other technologies to best benefit these communities. Project support from PanAsia This project will complement network infrastructure development being undertaken as part of a World Bank Human Resources Development project. In the World Bank project, TANUVAS will be able to provide the Internet connections to all 21 training and research centres operated by the University. A principal function of the system will be communication among centers and with the University for administrative purposes and for communication among scientists and specialists. This PanAsia project will add an extra dimension to that system by extending it to include communities and other local institutions. In summary, the PanAsia project will be concerned with needs analysis, models for making software user-friendly and relevant, and finding cultural appropriate means for providing community groups and individuals with direct and indirect access to the information technology and network resources at the centres. PROJECT OBJECTIVES The overall objective of this project is to develop and test tools that will support the effective use of information and communication technologies for rural development. Intermediate objectives include determining information needs of various population groups and designing systems for linking the groups to relevant information and raining resources. PROJECT BENEFICIARIES Principal project beneficiaries will be low income groups in the communities surrounding the three test centres. The project will focus especially on self-help womens groups, unemployed youth, and the small holder farmer population. To the extent that the results of the project are adopted in the University Training and Research Centres throughout the state, ultimately rural Tamil Nadu will be the beneficiary. PROJECT SUSTAINABILITY The Government of India has committed itself to the establishment of KVKs as a national priority. The UTRC system is firmly fixed in the structure of the University. Thus the institutions in which this project is spliced are on a firm foundation. The University will, in all likelihood, enter into a second phase of the World Bank Human Resources Development project which will mean continued support of the networking arrangements made in this first phase. TANUVAS itself has demonstrated the ability to innovate in related areas such as distance learning for farmers, and be creative in maintaining these initiatives through user fees and contracts. A strong theme in this research is the development of demand-driven services which will, as in the case of its distance learning

programmes, invite ways of gaining private and public sector support of the systems that emerge. PROJECT METHODS The project will build its program on ICT needs analysis using surveys, focus groups, and assessments of local leaders. It will organize a team of people within each of three centres and they will be responsible for working with specialists in designing the physical system that will provide members of the community with access to the centres IT. The teams will also collaborate to develop strategies for marketing the centres services and resources to target groups and to local institutions and NGOs. The teams will also work with University specialists and other organizations such as the Tamil Nadu Agricultural University and government bodies to build information resources. The project sites will include rural, peri-urban and urban settings. They will be at the Department of Communication and Entrepreneurship, Madras Veterinary College, Chennai (urban), Training Centre at Kattupakkam, Kanchipuram District (rural) and Veterinary College and research Institute at Namakkal (peri-urban). As additional resources become available from internal sources or other donors, more UTRCs will be incorporated into the research plan. PROJECT OUTPUTS The most significant project outputs will relate to improved welfare of the community. However, the more tangible outputs that can be clearly documented include the following: 1. A needs assessment tool tailored to information needs. This will be shared freely through professional meetings and www mechanisms. 2. An inventory of information needs in three communities, with particular sections devoted to special target groups. 3. A value-added process-model that will be shared freely through professional contacts. 4. A case study that will be relevant to telecenter developers in India elsewhere. 5. A collection of community-relevant software and a system for accessing software.

Project: INFORMATION TECHNOLOGY AND RURAL EXTENSION IN INDIA Collaborators: Tamil Nadu University of Veterinary and Animal Sciences Chennai, Tamil Nadu, India College of Agriculture and Life Sciences Cornell University Ithaca, New York USA Current status (2002): Telecenters have been opened in several locations. Community surveys focusing on information needs have been undertaken. PROJECT BACKGROUND AND JUSTIFICATION The focus of this project is on developing the software infrastructure to maximize the impact of a networked group of rural information centres in south India. The Tamil Nadu Veterinary and Animal Sciences University (TANUVAS) has a three part mission: (1) it conducts academic programmes for baccalaureate and post-graduate degree students; (2) it conducts research related to animal sciences and related areas such as food production and rural entrepreneurship; and (3) it carries out extension programmes for the benefit of rural dwellers, small businesses, cooperatives, various nongovernmental organizations, and government officials. This proposal is particularly related to the third part of the Universitys mission listed above: TANUVAS has a mandate to reach out to rural areas of Tamil Nadu to help communities and families improve their social and economic well-being. A substantial part of this extension effort is devoted to information and training programmes related to milk production, rabbitry, sheep and goat husbandry, small scale fisheries. One of the principal objectives of these efforts is to enable rural people to become small scale entrepreneurs and to insure that they have expert information support to help them sustain their enterprises. For example, the TANUVAS extension programme trains self-help rural womens groups in dairying, fisheries and other small-scale ventures, and provides information support to these groups through their group-elected representatives. Networking problem Currently TANUVAS tries to maintain contact with rural farmers and entrepreneurs through 14 University Training and Research Centres (UTRC), 3 Farmer Training Centres, and 2 Krishi Vigyan Kendras (KVK-Farm Science Centres) all scattered in rural towns throughout Tamil Nadu. These outreach services are staffed with animal and agricultural specialists, most of whom hold post graduate degrees and professional level rank in TANUVAS. The primary means of contact with rural constituents at the centers is

through face-to-face contact, either individually or in training programmes and workshops. While strategically placed to serve rural communities, the breadth of information services is largely limited to animal production issues. With the present system based heavily on labor-intensive personal contact, the centres do not reach a significant proportion of the population. Through a World Bank supported Human Resources Development Project, TANUVAS will connect each of the rural centers to the Internet. The linkage of these centres via modern information technologies to the University and to other expert resources will enable the centres to address more comprehensively and more quickly the information and communication needs of rural families. The intent of this project is to transform and institutionalize a group of these centres as Rural Information and Communication Centres (RICC). The Centres would expand the resources available to their constituent communities in two significant ways: (1) by making available a range of information and communication technologies (ICT) to the public; and (2) by increasing the breadth of information and training resources available, thus including current market data, family health and nutrition information, literacy materials, and e-commerce facilities. A major challenge is to link these centres with experts at TANUVAS using digital networks and thereby making more efficient the seeking and diffusion of information. Because of the added advantage of exchanges in this process, the information and other resources at the centres will be appropriate and relevant to the rural communities. There are important research dimensions in this project: 1. The first is to develop needs assessment tools specifically related to information, training, and communication technologies that can be used before start-up and continually thereafter so that the products and services of the centres are relevant to the variety of stakeholders in the communities. This means, for example, having a system that will find out the information and training needs of women and unemployed people who are often overlooked in development initiatives. The formulation, testing and refinement of these tools will provide models for expanding this project into more UTRCs and to telecentre-like projects elsewhere in the world. 2. The second research aspect is the development of methods and guidelines for training the community in the uses of the RICC, recognizing that the availability and accessability of such information centers does not automatically result in their use by the community and by community groups. For example, It will be a long, long time before all the obstacles to access are removed and everyone will voluntarily and spontaneously make use of the RICCs. It is important to recognize that while many in the community may themselves never use the centres directly, they may be able to benefit from them. Take, as an example, the more than a quarter of a million women who belong to self-help groups in Indias Tamil Nadu state. Telecenters could train representatives of these groups, who then would serve as liaisons between the groups and the RICCs. The liaison persons might also serve on an advisory committee for the centre, an important vehicle for

community participation in a centres operations. The results of this research will have relevance also for the thousands of telecentres being established in Asia, Africa and Latin America. 3. The third area of research deals with methods for adding value to information resources so that they will be valuable and user-friendly for the surrounding communities. There are many data bases and other information resources in India and across the world that potentially have relevance for Indian communities. For example, there are more than 12,000 www files related to health, but most of them are in the English language. An important dimension of this project is the role of the University as an information provider and value-adding hub for a rural communication network. In this case, the University, in its fundamental role of public service to the society, will be the key institution that will adapt demand?driven information to suit the needs of the communities served, providing information products that are accessible to the communities through the network of RICCs. To illustrate with another example, Infectious Bronchial Trachitis (IBR) in cattle and small ruminants is prevalent in both India and the USA. To control this, a vaccine F5 is being used in dairy farms in rural New York. The scientists at TANUVAS are continuously trying to isolate the virus and produce vaccine for IBR disease. No vaccine has yet been produced in India to control the resulting mastitis problem. Originating with Cornell University scientists, knowledge related to this problem can be widely shared among veterinary technical people and ultimately farmers if it can be processed through the value adding hub at TANUVAS. In this project we will test how universities can serve as support institutions that facilitate the introduction of Internet to rural and most disadvantaged areas. This project will experiment with how the Internet can be used and combined with other technologies to best benefit these communities. Project support from PanAsia This project will complement network infrastructure development being undertaken as part of a World Bank Human Resources Development project. In the World Bank project, TANUVAS will be able to provide the Internet connections to all 21 training and research centres operated by the University. A principal function of the system will be communication among centers and with the University for administrative purposes and for communication among scientists and specialists. This PanAsia project will add an extra dimension to that system by extending it to include communities and other local institutions. In summary, the PanAsia project will be concerned with needs analysis, models for making software user-friendly and relevant, and finding cultural appropriate means for providing community groups and individuals with direct and indirect access to the information technology and network resources at the centres. PROJECT OBJECTIVES The overall objective of this project is to develop and test tools that will support the effective use of information and communication technologies for rural development.

Intermediate objectives include determining information needs of various population groups and designing systems for linking the groups to relevant information and raining resources. PROJECT BENEFICIARIES Principal project beneficiaries will be low income groups in the communities surrounding the three test centres. The project will focus especially on self-help womens groups, unemployed youth, and the small holder farmer population. To the extent that the results of the project are adopted in the University Training and Research Centres throughout the state, ultimately rural Tamil Nadu will be the beneficiary. PROJECT SUSTAINABILITY The Government of India has committed itself to the establishment of KVKs as a national priority. The UTRC system is firmly fixed in the structure of the University. Thus the institutions in which this project is spliced are on a firm foundation. The University will, in all likelihood, enter into a second phase of the World Bank Human Resources Development project which will mean continued support of the networking arrangements made in this first phase. TANUVAS itself has demonstrated the ability to innovate in related areas such as distance learning for farmers, and be creative in maintaining these initiatives through user fees and contracts. A strong theme in this research is the development of demand-driven services which will, as in the case of its distance learning programmes, invite ways of gaining private and public sector support of the systems that emerge. PROJECT METHODS The project will build its program on ICT needs analysis using surveys, focus groups, and assessments of local leaders. It will organize a team of people within each of three centres and they will be responsible for working with specialists in designing the physical system that will provide members of the community with access to the centres IT. The teams will also collaborate to develop strategies for marketing the centres services and resources to target groups and to local institutions and NGOs. The teams will also work with University specialists and other organizations such as the Tamil Nadu Agricultural University and government bodies to build information resources. The project sites will include rural, peri-urban and urban settings. They will be at the Department of Communication and Entrepreneurship, Madras Veterinary College, Chennai (urban), Training Centre at Kattupakkam, Kanchipuram District (rural) and Veterinary College and research Institute at Namakkal (peri-urban). As additional resources become available from internal sources or other donors, more UTRCs will be incorporated into the research plan. PROJECT OUTPUTS

The most significant project outputs will relate to improved welfare of the community. However, the more tangible outputs that can be clearly documented include the following: 1. A needs assessment tool tailored to information needs. This will be shared freely through professional meetings and www mechanisms. 2. An inventory of information needs in three communities, with particular sections devoted to special target groups. 3. A value-added process-model that will be shared freely through professional contacts. 4. A case study that will be relevant to telecenter developers in India elsewhere. 5. A collection of community-relevant software and a system for accessing software.

Do Ethics Matter in Modern Journalism? Lt. Col. Rajan Suresh (Retd) (Lecturer, Department of Journalism, University of Kerala, Trivandrum, Kerala, India) Tehelka.com has been the most frequently discussed media entity in India for the past three weeks. Every day the media, politicians and bureaucrats have managed to place it on the public agenda one way or the other. But there seems to be no benefit in store for the billion plus Indian population. Judging by similar precedents, ordinary people have no tangibles to hope for. Tehelka did a very good job of duping some of the small fry in the arms business, and a few greedy civilian and army officers associated with arms purchases. They also managed to dupe important political functionaries such as Jaya Jaitley and R. K. Jain of the Samata Party and Bangaru Lakshman of the BJP. Tehelka established that many of their 'victims' took money, thereby going one step farther than Matt Drudge, the maverick American Journalist who set the trend of Web-journalism based mostly on sensational rumours. Tehelka obviously had their sights set on becoming 'top-dog' among the watchdogs of Indian polity. The reporter enjoyed celebrity status for a couple of days, and Zee TV, who had a stake in the sting, put the story out in all the channels they own, and most certainly raked in plenty of money. Tehelka's stated purpose- `to expose

corruption', has bolstered their image as a smart team of modern day "investigative" journalists. Although debates have ensued about the rights and wrongs in legal terms, the case has raised serious questions about journalistic ethics and responsibility. The West; particularly the United States and Britain, occasionally get to taste this genre of journalism. Private lives of `public figures' and private lives of private figures are often exposed by journalists looking for sensational stories. President Clinton and Princess Diana were the most prominent targets in recent times. Journalists can now become super peeping-toms with the help of hidden cameras and high-tech surveillance gadgets. And to absolve themselves of criminal intent, the sordid details of many such "investigations" are thrust on unsuspecting people, claiming that a public interest is being served. Western journalists also cheated and ruthlessly exploited ordinary people by impersonating social workers, counsellors or health-care personnel, to gather personal details that often led to painful consequences. A few classic examples can be found in Fineline, an American newsletter on Journalistic ethics. In a few cases, reporters managed to dupe the US Government, for getting first-hand information from protected sources. Robert Kapler, a small-time reporter pretending to be a security guard, breached the security set-up of a nuclear plant on Three Mile Island to prove that their security was weak. Jonathan Franklin, a freelance reporter signed up as a mortician at the Dover US Air Force Base to get accurate numbers of American soldiers who died in the Gulf War. Whether such reporting has brought about any change in the daily lives of American citizens, or whether it has done the public any good, are debatable. Tehelka used a wily combination of impersonation, hidden cameras and bribery (all these being culpable offences under Indian Law) to get their tape recordings. The Government of India contends that if public interest was a prime concern, the tapes should not have gone public in such an explosive fashion. But on the other hand, it can be argued quite logically that if the tapes had not been revealed so spectacularly, they would have lost much of their impact. No doubt the largest party in the ruling alliance has been shaken, and some of their political allies unnerved. The Defence Minister has resigned, and higher echelons of the armed forces have been tainted to some extent. India's opposition parties, led by the Congress (I) lost no time in calling for all sorts of "remedies". Tehelka has provided some political gains for the Congress (I). This leads us to the crux of the matter. What was the real purpose of this expose- Public interest or Tehelka's fame and fortune? Middlemen are undesirable elements in any kind of deal; but the reality is that deals don't go through without middlemen. Corruption is nothing new to the political and administrative set-up in independent India, and some members of the Armed Forces have committed crimes that are morally more serious; such as espionage for alien powers. The Samba spy scandal of the 70s and the later Coomer Narain- Larkins cases involved gratifications of far lesser magnitude for compromising national interests. But after nearly three decades, the incarcerated `spies' stand absolved of the crime, and we hear nothing about the Larkins. Wouldn't it be logical to assume that public attention would soon shift from the Tehelka expose? So the net effect would be to Tehelka's sole advantage, unless

Tarun Tejpal and his team persevere, forcing the judiciary or Parliament to set wrongs right. Will they keep at it, or withdraw from the scene saying that their job is over? Tehelka has broken the story with damning footage on video tape, albeit a little out of focus and frequently out of perspective. The soundtracks and subtitles took on the task of convincing viewers about deals being struck and money changing hands. Unfortunately the odds do not favour Tehelka, simply because modern technology offers all and sundry the wherewithal to digitally manipulate moving images and sound through easy-to-use software packages. It would be a fairly simple task to `smart-edit' video footage, add doctored sound tracks and put all sorts of filters to blur the final product so that it appears authentic. This is not an allegation that Tehelka has done so, but it would take a lot more proof before knowledgeable people are convinced of the video's genuineness. Spielberg's Jurassic Park has made us all skeptical about such `reality'. Perhaps the best thing the Tehelka tapes did was to highlight human greed. They brought out a story that actually sent ripples through the Government, and made people sit up and take notice. They managed to assign values to individual gullibility. It is now established that the ruling party's titular head can be bought for a hundred thousand rupees and a Major General in the Army for a fifth of that. There is no need to discuss the pros and cons of the matter or to probe whether the money went to party coffers or individuals' pockets. And there is no justification in saying that "money was taken because it was offered". Public figures have misbehaved, and Tehelka's reporters have exposed them. Tehelka.com, on their part, have taken the law into their hands, and played a con game on unsuspecting (though mostly corrupt) persons. It does not constitute normal journalistic behaviour by any standards. A question of whether this investigative exercise has caused harm to the public is also being discussed, because speculations are rife that Indian soldiers as a whole are demoralized and feel `let down' by the alleged compromises in arming and equipping them. If this situation is for real, then the consequences are not so much in the `public interest', and there may be reason for the Government to brand Tejpal, Bahl and Mathew as mercenaries and traitors. It is catch 22 for the trio. But since we have Dawood Ibrahim, Ottavio Quattrocchi, Harshad Mehta and the more down-toearth Veerappan calling the shots, it seems anyone can get away with anything in India. Tehelka.com ought to survive, and continue entertaining the Indian public in true Drudge fashion. The ethical issues in this case remain anchored on two aspects. One being the motivation for Tehelka.com to take up the issue of bribes and commissions in defence deals at this point of time, when far bigger cases such as Bofors , HDW and Tangushka are still unresolved. It would do well for Tejpal to explain whether it was journalistic vigilantism or self-interest that encouraged Tehelka reporters to give away large amounts of cash and take such risks. The other pertains to the method adopted- in befriending touts, feeding, wining and bribing them, and on false pretexts, recording their tall claims with hidden cameras, and finally offering these recordings as news to the public. There is a need to supplement the sensational footage with more real proof and explanation; particularly to ensure that the `public' are not fed half-truths and speculations.